Rehabilitation and Parkinson's Disease

Parkinson’s Disease

Guest Editors: Gammon M. Earhart, Lee Dibble, Terry Ellis, and Alice Nieuwboer

Rehabilitation and Parkinson’s Disease


Parkinson’s Disease


Guest Editors: Gammon M. Earhart, Terry Ellis,Alice Nieuwboer, and Leland E. Dibble

Copyright © 2012 Hindawi Publishing Corporation. All rights reserved.

This is a special issue published in “Parkinson’s Disease.” All articles are open access articles distributed under the Creative CommonsAttribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is prop-erly cited.

Editorial Board

Jan O. Aasly, NorwayCristine Alves da Costa, FranceIvan Bodis-Wollner, USAD. J. Brooks, UKCarlo Colosimo, ItalyMark R. Cookson, USAAlan R. Crossman, UKT. M. Dawson, USAH. J. Federoff, USA

Francisco Grandas, SpainPeter Hagell, SwedenN. Hattori, JapanMarjan Jahanshahi, UKE. D. Louis, USAP. Martinez Martin, SpainF. Mastaglia, AustraliaHuw R. Morris, UKM. Maral Mouradian, USA

Antonio Pisani, ItalyJose Rabey, IsraelHeinz Reichmann, GermanyFabrizio Stocchi, ItalyEng King Tan, SingaporeHelio Teive, BrazilDaniel Truong, USAYoshikazu Ugawa, Japan

Contents

Rehabilitation and Parkinson’s Disease, Gammon M. Earhart, Terry Ellis, Alice Nieuwboer,and Leland E. DibbleVolume 2012, Article ID 371406, 3 pages

LSVT LOUD and LSVT BIG: Behavioral Treatment Programs for Speech and Body Movement inParkinson Disease, Cynthia Fox, Georg Ebersbach, Lorraine Ramig, and Shimon SapirVolume 2012, Article ID 391946, 12 pages

Improving Community Healthcare for Patients with Parkinson’s Disease: The Dutch Model, S. H. J. Keus,L. B. Oude Nijhuis, M.J. Nijkrake, B. R. Bloem, and M. MunnekeVolume 2012, Article ID 543426, 7 pages

Altered Dynamic Postural Control during Step Turning in Persons with Early-Stage Parkinsons Disease,Jooeun Song, Susan Sigward, Beth Fisher, and George J. SalemVolume 2012, Article ID 386962, 8 pages

Is Freezing of Gait in Parkinson’s Disease a Result of Multiple Gait Impairments? Implicationsfor Treatment, Meir Plotnik, Nir Giladi, and Jeffrey M. Hausdorff

Volume 2012, Article ID 459321, 8 pages

Posture and Locomotion Coupling: A Target for Rehabilitation Interventions in Persons withParkinson’s Disease, Marie-Laure Mille, Robert A. Creath, Michelle G. Prettyman,Marjorie Johnson Hilliard, Katherine M. Martinez, Colum D. MacKinnon, and Mark W. RogersVolume 2012, Article ID 754186, 10 pages

Feasibility, Safety, and Compliance in a Randomized Controlled Trial of Physical Therapy for Parkinson’sDisease, Jennifer L. McGinley, Clarissa Martin, Frances E. Huxham, Hylton B. Menz, Mary Danoudis,Anna T. Murphy, Jennifer J. Watts, Robert Iansek, and Meg E. MorrisVolume 2012, Article ID 795294, 8 pages

Improved Dynamic Postural Task Performance without Improvements in Postural Responses:The Blessing and the Curse of Dopamine Replacement, K. B. Foreman, C. Wisted, O. Addison,R. L. Marcus, P. C. LaStayo, and L. E. DibbleVolume 2012, Article ID 692150, 8 pages

Upper Extremity Motor Learning among Individuals with Parkinson’s Disease: A Meta-AnalysisEvaluating Movement Time in Simple Tasks, K. Felix, K. Gain, E. Paiva, K. Whitney, M. E. Jenkins,and S. J. SpauldingVolume 2012, Article ID 589152, 7 pages

Lack of Short-Term Effectiveness of Rotating Treadmill Training on Turning in People withMild-to-Moderate Parkinson’s Disease and Healthy Older Adults: A Randomized, Controlled Study,Marie E. McNeely and Gammon M. EarhartVolume 2012, Article ID 623985, 8 pages

Accuracy of Fall Prediction in Parkinson Disease: Six-Month and 12-Month Prospective Analyses,Ryan P. Duncan, Abigail L. Leddy, James T. Cavanaugh, Leland E. Dibble, Terry D. Ellis, Matthew P. Ford,K. Bo Foreman, and Gammon M. EarhartVolume 2012, Article ID 237673, 7 pages

Community Walking in People with Parkinson’s Disease, Robyn M. Lamont, Meg E. Morris,Marjorie H. Woollacott, and Sandra G. BrauerVolume 2012, Article ID 856237, 8 pages

Progressive Resistance Exercise and Parkinson’s Disease: A Review of Potential Mechanisms,Fabian J. David, Miriam R. Rafferty, Julie A. Robichaud, Janey Prodoehl, Wendy M. Kohrt,David E. Vaillancourt, and Daniel M. CorcosVolume 2012, Article ID 124527, 10 pages

Exercise and Motor Training in People with Parkinson’s Disease: A Systematic Review of ParticipantCharacteristics, Intervention Delivery, Retention Rates, Adherence, and Adverse Events in Clinical Trials,Natalie E. Allen, Catherine Sherrington, Gayanthi D. Suriyarachchi, Serene S. Paul, Jooeun Song,and Colleen G. CanningVolume 2012, Article ID 854328, 15 pages

Effectiveness of an Inpatient Movement Disorders Program for Patients with Atypical Parkinsonism,Anna D. Hohler, Jyeming M. Tsao, Douglas I. Katz, T. Joy DiPiero, Christina L. Hehl, Alissa Leonard,Valerie Allen, Maura Gardner, Heidi Phenix, Marie Saint-Hilaire, and Terry EllisVolume 2012, Article ID 871974, 6 pages

A Review of Dual-Task Walking Deficits in People with Parkinson’s Disease: Motor and CognitiveContributions, Mechanisms, and Clinical Implications, Valerie E. Kelly, Alexis J. Eusterbrock,and Anne Shumway-CookVolume 2012, Article ID 918719, 14 pages

Reliability in One-Repetition Maximum Performance in People with Parkinson’s Disease,Thomas A. Buckley and Christopher J. HassVolume 2012, Article ID 928736, 6 pages

Comparing the Mini-BESTest with the Berg Balance Scale to Evaluate Balance Disorders in Parkinson’sDisease, Laurie A. King, Kelsey C. Priest, Arash Salarian, Don Pierce, and Fay B. HorakVolume 2012, Article ID 375419, 7 pages

The PIT: SToPP TrialA Feasibility Randomised Controlled Trial of Home-Based Physiotherapy forPeople with Parkinson’s Disease Using Video-Based Measures to Preserve Assessor Blinding,Emma Stack, Helen Roberts, and Ann AshburnVolume 2012, Article ID 360231, 8 pages

Gait Difficulty, Postural Instability, and Muscle Weakness Are Associated with Fear of Falling in Peoplewith Parkinson’s Disease, Margaret K. Y. Mak, Marco Y. C. Pang, and Vincent MokVolume 2012, Article ID 901721, 5 pages

A Manipulation of Visual Feedback during Gait Training in Parkinson’s Disease, Quincy J. Almeida andHaseel BhattVolume 2012, Article ID 508720, 7 pages

Walking Ability Is a Major Contributor to Fear of Falling in People with Parkinson’s Disease:Implications for Rehabilitation, Maria H. Nilsson, Gun-Marie Hariz, Susanne Iwarsson, and Peter HagellVolume 2012, Article ID 713236, 7 pages

Impaired Economy of Gait and Decreased Six-Minute Walk Distance in Parkinson’s Disease,Leslie I. Katzel, Frederick M. Ivey, John D. Sorkin, Richard F. Macko, Barbara Smith, and Lisa M. ShulmanVolume 2012, Article ID 241754, 6 pages

Hindawi Publishing CorporationParkinson’s DiseaseVolume 2012, Article ID 371406, 3 pagesdoi:10.1155/2012/371406

Editorial


Gammon M. Earhart,1 Terry Ellis,2 Alice Nieuwboer,3 and Leland E. Dibble4

1 Program in Physical Therapy, Department of Anatomy and Neurobiology and Department of Neurology,Washington University in St. Louis School of Medicine, St. Louis, MO 63108, USA

2 Department of Physical Therapy and Athletic Training, Boston University, Boston, MA 02215, USA3 Department of Rehabilitation Sciences, Katholieke Universiteit Leuven, 3001 Leuven, Belgium4 Department of Physical Therapy, University of Utah, Salt Lake City, UT 84108, USA

Correspondence should be addressed to Gammon M. Earhart, [email protected]

Received 14 December 2011; Accepted 14 December 2011

Copyright © 2012 Gammon M. Earhart et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

Early after the turn of the century, much excitement wasgenerated by the reports of Tillerson et al. [1, 2] thatexercise appeared to protect against neuronal degenerationin rodent models of toxin-induced parkinsonism. Such find-ings, coupled with epidemiologic suggestions that personswith a history of moderate to vigorous exercise may have adecreased risk of developing Parkinson’s disease (PD) [3–5],led to an exponential growth in research on the effects ofphysical activity and exercise on PD. Unfortunately, addi-tional follow-up animal studies to the work of Tillerson et al.have failed to yield consistent findings [6–10]. For thisreason, it appears that the critical factors associated withneuroprotection remain elusive. With a continued focuson examining the effects of exercise in animal models ofparkinsonism, identifying biomarkers of disease progression,and new and innovative outcomes, we look forward to aday when an evidence-based neuroprotection study can beimplemented in human idiopathic PD.

Although results from studies of the neuroprotectiveeffects of exercise are mixed, one consistent finding fromanimal models and human trials is the lack of adverse effectsof exercise and physical activity on anatomic and behavioraloutcomes. The adverse side-effect profile of exercise as anintervention for those with PD appears to be minimal. Assuch, we think there is no reason to wait for confirmationof neuroprotection. Rather, evidence is accumulating thatexercise and physical activity should be utilized as key toolsin the management of PD across the spectrum of disease.Evidence-based approaches to rehabilitation are known to

improve physical functioning, strength, balance, gait, andhealth-related quality of life among people with PD [11–13],but questions remain about whether or not these approachescan substantially impact fall rates [14–16]. This is a keyissue, as most individuals with PD are only referred torehabilitation after the onset of reduced mobility and anincrease in falls. As such, the majority of PD rehabilitationcare is provided in a tertiary prevention model of care.People with PD are most often not seen earlier in the courseof the disease, when rehabilitation could play a key rolein secondary preventive care. Secondary prevention wouldentail addressing early PD signs and symptoms, ideallyimmediately upon diagnosis, to optimize the condition of thecentral nervous system as well as other peripheral systemssuch as the cardiovascular and respiratory systems in orderto maximize function and slow progression of disability.Even earlier intervention should be considered, as we thinkthat rehabilitation may ultimately serve a role in primaryprevention of PD. Primary prevention would entail treatingthose without current neurologic signs and symptoms inorder to prevent PD from ever developing. Those who arepotentially at risk for PD (e.g., leucine-rich repeat kinase2 (LRRK2) carriers, those with rapid eye movement sleepbehavior disorder, anosmia, constipation, abnormal positronemission tomography (PET) scans, etc.) may be excellentcandidates for primary prevention.

The presently limited scope of rehabilitation in themanagement of PD, with utilization of rehabilitation asmainly a tertiary prevention measure, reflects a missed

2 Parkinson’s Disease

opportunity on the part of healthcare providers, patientadvocacy groups, and patients themselves. All stakeholdersin this situation should be advocates for higher expectationsand should work together to develop targets for the futureof rehabilitation in PD to include primary and secondaryprevention and to improve our current provision of tertiaryprevention interventions.

From the Spectrum of Rehabilitation Options to AtypicalParkinsonism. In this special issue are several articles that wehope advance the field and move us closer to these futuretargets. The issue opens with a series of three review articles.The first paper summarizes and synthesizes the nature andfeatures of previous randomized, controlled trials of exerciseor motor training in PD, important for the long-termprovision of increasing levels of physical activity. The secondpaper provides a meta-analysis focused on motor learning inupper extremity tasks. The third paper provides an integratedoverview of the Lee Silverman Voice Treatment approachto voice and movement therapy, discussing the rationalefor the approach as well as the data regarding efficacy.These opening three articles highlight the important roles ofspeech, occupational, and physical therapy approaches in therehabilitation of individuals with PD as well as address areasfor future research.

Several papers in this special issue relate to gait, balance,and falls, examining the relationship between gait economyand six-minute walk distance, reviewing the literature on thecosts of dual-task walking, and providing a new theoreticalframework for considering freezing of gait (FOG) includingmethods to improve overall locomotor performance andmethods to target the triggers of FOG. Also included are tworandomized, controlled trials designed to improve walking.One compares visually cued walking training on a treadmillto overground walking; the other examines use of rotatingtreadmill training as a means of improving turning, therebytargeting a known trigger of FOG. Turning is also examinedin a paper that describes turning impairments in early PD.This is followed by a paper examining the effects of medica-tions on gait-related mobility and postural control, showingthat although pharmacologic intervention enhanced someaspects of mobility, reactive postural responses did notimprove. This highlights the need for awareness of posturalcontrol deficits and the need to be able to measure thesedeficits, as is addressed by the next two papers in the issuethat present the relative merits of different balance measuresacross different levels of PD severity and examine the relativeeffectiveness of different balance measures for prospective fallprediction.

The inextricable link between posture and gait isaddressed in a paper that focuses on the coupling of postureand locomotion and suggests this coupling as a specifictarget for rehabilitation. This link is also highlighted intwo papers demonstrating that walking ability is a majorcontributor to fear of falling in PD, as is knee strength. Thelatter leads to the suggestion that resistance training maytherefore be warranted as an approach to reduce fear offalling. The rationale for progressive resistance training aswell as its potential mechanisms are addressed in a review

paper, followed by an article examining the reliability of onerepetition maximum strength testing in PD.

We conclude the issue with a set of papers addressingthe delivery of evidence-based rehabilitation in differentsettings. These papers include two that are randomized,controlled trials of physiotherapy in outpatient and home-based settings, respectively. These are followed by a paperexamining facilitators and barriers to community-basedwalking exercise among those with PD. Community-basedhealthcare for PD is also addressed in a paper describingthe steps taken to improve the Dutch model of multidisci-plinary care. The final paper of the issue demonstrates theeffectiveness of multidisciplinary care in an inpatient settingin persons with atypical parkinsonism. All papers in thisfinal section draw attention to the need for a collaborative,cooperative approach to rehabilitation across disciplines,across settings, and with PD and other related disorders.

Ultimately, we believe that there is a need to redefinethe role of rehabilitation in PD to include the provisionof primary, secondary, and tertiary prevention approaches.Across this spectrum from primary through tertiary care,the application of multidisciplinary approaches is neededto optimize the health, function, and quality of life ofindividuals at risk for, or who already have, PD. Only thenwill the full potential of rehabilitation in the management ofPD be realized.

Gammon M. EarhartTerry Ellis

Alice NieuwboerLeland E. Dibble

References

[1] J. L. Tillerson, A. D. Cohen, W. M. Caudle, M. J. Zigmond,T. Schallert, and G. W. Miller, “Forced nonuse in unilateralParkinsonian rats exacerbates injury,” Journal of Neuroscience,vol. 22, no. 15, pp. 6790–6799, 2002.

[2] J. L. Tillerson, A. D. Cohen, J. Philhower, G. W. Miller, M.J. Zigmond, and T. Schallert, “Forced limb-use effects on thebehavioral and neurochemical effects of 6-hydroxydopamine,”Journal of Neuroscience, vol. 21, no. 12, pp. 4427–4435, 2001.

[3] M. Hamer and Y. Chida, “Physical activity and risk ofneurodegenerative disease: a systematic review of prospectiveevidence,” Psychological Medicine, vol. 39, no. 1, pp. 3–11,2009.

[4] E. L. Thacker, H. Chen, A. V. Patel et al., “Recreational physicalactivity and risk of Parkinson’s disease,” Movement Disorders,vol. 23, no. 1, pp. 69–74, 2008.

[5] Q. Xu, Y. Park, X. Huang et al., “Physical activities and futurerisk of Parkinson disease,” Neurology, vol. 75, no. 4, pp. 341–348, 2010.

[6] M. Al-Jarrah, K. Pothakos, L. Novikova et al., “Enduranceexercise promotes cardiorespiratory rehabilitation withoutneurorestoration in the chronic mouse model of Parkinsonismwith severe neurodegeneration,” Neuroscience, vol. 149, no. 1,pp. 28–37, 2007.

[7] M. Mabandla, L. Kellaway, A. S. C. Gibson, and V. A. Russell,“Voluntary running provides neuroprotection in rats after 6-hydroxydopamine injection into the medial forebrain bundle,”Metabolic Brain Disease, vol. 19, no. 1-2, pp. 43–50, 2004.

Parkinson’s Disease 3

[8] M. V. Mabandla and V. A. Russell, “Voluntary exercise reducesthe neurotoxic effects of 6-hydroxydopamine in maternallyseparated rats,” Behavioural Brain Research, vol. 211, no. 1, pp.16–22, 2010.

[9] S. J. O’Dell, N. B. Gross, A. N. Fricks, B. D. Casiano, T. B.Nguyen, and J. F. Marshall, “Running wheel exercise enhancesrecovery from nigrostriatal dopamine injury without inducingneuroprotection,” Neuroscience, vol. 144, no. 3, pp. 1141–1151,2007.

[10] M. C. Yoon, M. S. Shin, T. S. Kim et al., “Treadmill exer-cise suppresses nigrostriatal dopaminergic neuronal loss in6-hydroxydopamine-induced Parkinson’s rats,” NeuroscienceLetters, vol. 423, no. 1, pp. 12–17, 2007.

[11] M. J. Falvo, B. K. Schilling, and G. M. Earhart, “Parkinson’sdisease and resistive exercise: rationale, review, and recom-mendations,” Movement Disorders, vol. 23, no. 1, pp. 1–11,2008.

[12] V. A. Goodwin, S. H. Richards, R. S. Taylor, A. H. Taylor, andJ. L. Campbell, “The effectiveness of exercise interventions forpeople with Parkinson’s disease: a systematic review and meta-analysis,” Movement Disorders, vol. 23, no. 5, pp. 631–640,2008.

[13] S. H. J. Keus, M. Munneke, M. J. Nijkrake, G. Kwakkel, and B.R. Bloem, “Physical therapy in Parkinson’s disease: evolutionand future challenges,” Movement Disorders, vol. 24, no. 1, pp.1–14, 2009.

[14] N. E. Allen, C. G. Canning, C. Sherrington et al., “The effectsof an exercise program on fall risk factors in people withParkinson’s disease: a randomized controlled trial,” MovementDisorders, vol. 25, no. 9, pp. 1217–1225, 2010.

[15] N. E. Allen, C. Sherrington, S. S. Paul, and C. G. Canning,“Balance and falls in Parkinson’s disease: a meta-analysis ofthe effect of exercise and motor training,” Movement Disorders,vol. 26, no. 9, pp. 1605–1615, 2011.

[16] V. A. Goodwin, S. H. Richards, W. Henley, P. Ewings, A. H.Taylor, and J. L. Campbell, “An exercise intervention to preventfalls in people with Parkinson’s disease: a pragmatic ran-domised controlled trial,” Journal of Neurology, Neurosurgeryand Psychiatry, vol. 82, no. 11, pp. 1232–1238, 2011.


Review Article

LSVT LOUD and LSVT BIG: Behavioral Treatment Programs forSpeech and Body Movement in Parkinson Disease

Cynthia Fox,1 Georg Ebersbach,2 Lorraine Ramig,1 and Shimon Sapir3

1 National Center for Voice and Speech, University of Colorado Boulder, Campus Box 409, Boulder, CO 80305, USA2 Movement Disorders Clinic, Paracelsusring 6a, 14547 Beelitz-Heilstatten, Germany3 Departments of Physiotherapy and Communication Sciences and Disorders, University of Haifa, Mount Carmel, Haifa 31905, Israel

Correspondence should be addressed to Lorraine Ramig, [email protected]

Received 2 August 2011; Revised 2 November 2011; Accepted 6 November 2011

Academic Editor: Leland E. Dibble

Copyright © 2012 Cynthia Fox et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Recent advances in neuroscience have suggested that exercise-based behavioral treatments may improve function and possibly slowprogression of motor symptoms in individuals with Parkinson disease (PD). The LSVT (Lee Silverman Voice Treatment) Programsfor individuals with PD have been developed and researched over the past 20 years beginning with a focus on the speech motorsystem (LSVT LOUD) and more recently have been extended to address limb motor systems (LSVT BIG). The unique aspects of theLSVT Programs include the combination of (a) an exclusive target on increasing amplitude (loudness in the speech motor system;bigger movements in the limb motor system), (b) a focus on sensory recalibration to help patients recognize that movements withincreased amplitude are within normal limits, even if they feel “too loud” or “too big,” and (c) training self-cueing and attentionto action to facilitate long-term maintenance of treatment outcomes. In addition, the intensive mode of delivery is consistent withprinciples that drive activity-dependent neuroplasticity and motor learning. The purpose of this paper is to provide an integrativediscussion of the LSVT Programs including the rationale for their fundamentals, a summary of efficacy data, and a discussion oflimitations and future directions for research.

1. Introduction

Progressive neurological diseases, such as Parkinson disease(PD) impair speech, swallowing, limb function, gait, balance,and activities of daily living. Even with optimal medical man-agement (pharmacological, surgical) these deficits cannot becontrolled satisfactorily in the vast majority of individualswith PD and have a negative impact on quality of life [1–3]. Recently, basic science research in animal models of PDhas documented the value of exercise for improving motorperformance and potentially slowing progression of motorsymptoms and neural degeneration [4–9]. The impact ofexercise in humans with PD is being increasingly exploredin studies that incorporate key principles that have beenidentified to drive activity-dependent neuroplasticity (i.e.,modifications in the central nervous system in response tophysical activity), such as specificity, intensity, repetition,and saliency [9–16]. Collectively, these findings have accen-tuated the important role of exercise and/or rehabilitation

in the overall management of PD. Previously, rehabilitationprograms were often administered in later stages of PDor as reactive referrals for secondary impairments, such asaspiration due to swallowing dysfunction, or hip fracturedue to falling. Today, such programs are being viewed astherapeutic options to be prescribed early in the course ofPD that may potentially contribute to slowing of motorsymptom progression [5, 17]. The purpose of this paperis to provide an integrative discussion of the rationale forand the efficacy of one type of rehabilitation approach,the LSVT Programs for speech (LSVT LOUD) and limb(LSVT BIG) motor systems in individuals with PD. We willinclude the rationale for targeting increased amplitude, theintensive mode of treatment delivery, and recalibration ofthe sensorimotor system including self-cueing, and attentionto action, which may be important for generalization andlong-term maintenance of treatment effects. In addition, wewill summarize published efficacy data and discuss currentlimitations and future directions for research.


2. What Is LSVT LOUD?

Nearly 90% of individuals with PD have speech and voicedisorders that negatively impact communication abilities[18, 19]. These disorders include reduced vocal loudness,monotone, hoarse, breathy voice quality, and imprecisearticulation, perceived as mumbling, and other rate-relatedfeatures, such as hesitations and short rushes of speech[20, 21]. In contrast to previous medical “chart review”literature suggesting a mid- or late-stage onset of speech andswallowing symptoms in PD [22], more recent investigationswith sensitive and valid measures consistently report speechsymptoms in early PD (e.g., [23]). Further, self-report datafrom individuals with PD have indicated that voice andspeech changes are associated with inactivity, embarrass-ment, and withdrawal from social situations [2].

Historically, speech treatment for individuals with PDwas viewed as futile, in as much as treatment gains wereminimal and short lived [24]. Today, LSVT LOUD is astandardized, research-based speech treatment protocol withestablished efficacy for PD [25–28]. LSVT LOUD trainsthe target of vocal loudness in order to (1) enhance thevoice source, consistent with improving the carrier in theclassic engineering concept of signal transmission [29], (2)use vocal loudness as a trigger for distributed effects (e.g.,improved articulation, vocal quality and intonation, andreduced rate) across the speech production system [21, 30–33], (3) recalibrate sensorimotor perception of improvedvocal loudness [34], and (4) train a single self-cue andattention to action to facilitate generalization of treatmenteffects into functional communication. Although LSVTLOUD is a standardized treatment protocol, the materialsused during treatment and the homework and carryoverexercises are made salient and tailored to each individual tofacilitate motivation, engagement and the potential to driveneuroplasticity [13, 35, 36].

In contrast, traditional speech therapy typically involvesmultiple speech system targets (e.g., respiration, voice,articulation, and rate), is low intensity (1–2 sessions perweek, minimal number of repetitions of treatment tasks),and does not systematically address the sensory processingdeficits related to self-perception of loudness by individualswith PD (see [37] for summary table contrasting LSVTLOUD and traditional speech treatment) [37, 38]. The LSVTLOUD protocol is summarized in Table 1.

3. LSVT LOUD Outcome Data

Two randomized controlled trial (RCT) studies have beenconducted [27, 28]. Data have documented that trainingincreased vocal loudness results in a statistically significantand lasting increase in vocal sound pressure level (SPL) andfrequency variability during speech (i.e., uncued conversa-tional speech) as compared to a matched treatment focusingon training increased respiratory support [26–28, 30]. Effectsize data for the primary outcome variable of vocal SPL inconversational speech were highly significant immediatelyposttreatment (1.20) and were maintained at 24 monthsposttreatment (1.03) [27, 30, 40]. Data providing initial

external validation of LSVT LOUD outcomes have beenreported by independent labs and reviews [41–45].

In addition, various physiologic changes such as in-creased movement amplitude of the rib cage (larger excur-sions) during speech breathing [46], increased subglottalair pressure [26], and improved closure and larger/moresymmetrical movements of the vocal folds [47] have beendocumented in individuals with PD immediately after LSVTLOUD. These findings are supported by perceptual datademonstrating listeners rated improved loudness and voicequality in individuals with PD immediately posttreatment[33]. Subjects in these studies were predominately Hoehnand Yahr stages 1–3 with moderate speech deficits.

Training vocal loudness also has been studied for itsdistributed effects across the speech production system. Ina series of smaller pilot studies (subsets of data from largerstudy) data have documented improvements in orofacialmovements, as reflected in consonant articulation [48],tongue strength and motility [44], speech rate [30], ratingsof improved facial expression [49], and improvements insome aspects of the oral phase of swallowing (e.g., reducedoral transit time) [50] even though these functions werenot specific targets in therapy. The impact of LSVT LOUDon speech articulation, especially vowels, has been furtherexplored. Vowels are formed and differentiated from eachother by the movements of the tongue, lips, and jaw. Inindividuals with PD, these movements tend to be hypokinetic[51], thus rendering the vowels less distinct physiologically,acoustically, and perceptually, a phenomenon known asvowel centralization. LSVT LOUD has been shown toreduce vowel centralization and improve perceptual ratingof vowel quality [31, 32]. This improvement may reflectlarger amplitude of movements of the tongue, lips, and jaw,possibly due to overall neural and biomechanical couplingof speech subsystems and increased activation of the entirespeech neuromuscular system [52].

Two brain imaging studies using O15 PET in a smallnumber of individuals with PD have documented changesin brain function immediately following LSVT LOUD[53–55]. The most recent study by Narayana et al. [55]examined the neural mechanisms underlying the effectsof training increased vocal loudness in ten individualswith PD and hypophonia. Cerebral blood flow during restand reading conditions was measured by H2

15O-positronemission tomography. Z-score images were generated bycontrasting reading with rest conditions for pre- and post-LSVT LOUD sessions, and neural activity was correlatedwith the corresponding change in vocal SPL (loudness).Narayana et al. [55] hypothesized that brain activationpatterns associated with LSVT LOUD training would reflectimproved loudness, improved perception of self-generatedvoice output, and improved attention to action. Furtherit was hypothesized that these outcomes would likely bemediated via the right hemisphere and involve speech motorand premotor cortical areas (related to increasing vocalloudness), the auditory cortices (related to recalibrationof perception of self-produced loudness), and dorsolateralprefrontal cortex (related to improving attention to action).To a large extent, the results of the study are consistent with


Table 1: Comparison of LSVT LOUD and LSVT BIG treatments.

LSVT LOUD (e.g., [25, 30]) LSVT BIG (e.g., [39])

Target: LOUD Target: BIG

Increased movement amplitude directed predominately torespiratory/laryngeal systems

Increased movement amplitude directed across limb motorsystem including gait

Intensity: standardized Intensity: standardized

Dosage: 4 consecutive days a week for 4 weeks (16 sessions inone month)Repetitions: minimum 15 repetitions/taskEffort: push for maximum patient-perceived effort each day(8 or 9 on scale of 1–10 with 10 being the most)

Dosage: 4 consecutive days a week for 4 weeks (16 sessions inone month)Repetitions: minimum 8–16 repetitions/taskEffort: push for maximum patient-perceived effort each day(8 or 9 on scale of 1–10 with 10 being the most)

Daily exercises Daily exercises

First half of the treatment session (30 min.)Task 1: Maximum Sustained Movements15 reps: sustain “ah” in Loud good quality voice as long aspossibleTask 2: Directional Movements15 reps each: say “ah” in Loud good quality voice going highin pitch;15 reps each: say “ah” in Loud good quality voice going lowin pitchTask 3: Functional PhrasesPatient self-identifies 10 phrases or sentences he/she saysdaily in functional living (e.g., “Good morning”)5 reps of the list of 10 phrases. “Read phrases using sameeffort/loudness as you did during the long “ah”

First half of the treatment session (30 min. or more)Task 1: Maximum Sustained Movements: seated8 reps: sustain Big “stretch” floor to ceiling (10 sec hold);8 reps: sustain Big “stretch” side to side (10 sec hold)Task 2: Repetitive/Directional Movements: standing16 reps: Forward Big step – 8 each leg;16 reps: Sideways Big step – 8 each side;16 reps: Backward Big step – 8 each leg;20 reps: Forward Big Rock and reach – 10 each side;20 reps: Sideways Big Rock and reach – 10 each sideTask 3: Functional Component MovementsPatient self-identifies 5 movements he/she does in functionalliving every day (e.g., Sit-to-stand)Clinician and patient select one simple component of each ofthese movements5 reps each of the 5 component movements “Do yourmovement with the same effort/bigness that you did duringthe daily exercises”

Hierarchy Hierarchy

Second half of the treatment session (30 min)(i) Designed to train rescaled amplitude/effort of movementachieved in daily exercises and functional phrases into incontext specific and variable speaking activities(ii) Tasks increase complexity across weeks(Words-phrases-sentences-reading-conversation) and can betailored to each patient’s goals and interests (e.g., golf versuscooking)(iii) Tasks progress in difficulty by increasing duration(maintain LOUD for longer periods of time) amplitude(loudness, within normal limits), and complexity of tasks(dual processing, background noise, and attentionaldistracters)

Second half of the treatment session (30 min or less)(i) Designed to train rescaled amplitude/effort of movementachieved in daily exercises and functional componentmovements into in context specific and variable movementactivities(ii) Complex multilevel tasks that progressively become moredifficult over the 4 weeks and can be tailored to each patient’sgoals and interests (e.g., basic bathroom skills versus goingout to dinner or shopping)(iii) Tasks progress in difficulty by increasing duration(maintain BIG for longer periods of time) amplitude(bigness/effort, within normal limits), and complexity oftasks (multisteps, dual processing, background noise, andattentional distracters)(iv) BIG walking is included as part of hierarchy on a dailybasis. Time and distance will vary across patients, hierarchygoals, and weeks of therapy

Shaping techniques Shaping techniques

Goal: train vocal loudness that is healthy and good quality(i.e., no unwanted vocal strain or excessive vocal fold closure)Technique: shape the quality and voice loudness through useof modeling or tactile/visual cues. “Watch me and do what Ido.”Minimal cognitive loading: behavior is not achieved throughextensive instructions or explanations, which are often toocomplex for patient to generalize outside of treatment room,but rather the patient is trained through modeling

Goal: train movement bigness that is healthy and goodquality (i.e., no unwanted strain or pain, impingement, orawkward biomechanics)Technique: shape the quality and movement bigness throughuse of modeling or tactile/visual cues. “Watch me and dowhat I do.”Minimal cognitive loading: behavior is not achieved throughextensive instructions or explanations, which are often toocomplex for patient to generalize outside of treatment room,but rather the patient is trained through modeling


Table 1: Continued.

LSVT LOUD (e.g., [25, 30]) LSVT BIG (e.g., [39])

Sensory recalibration Sensory recalibration

Treatment: focus attention on how it feels and sounds to talkLOUDCarryover activities: start day one; daily assignments(treatment and nontreatment days); Loud good quality voicein real-life situations; (i) difficulty of the assignment matchesthe level of the hierarchy where the person is working; (ii)make patient accountable and probe for comments frompatient that people in their daily living have said, such as, “Ican hear you better”Homework practice: start day one: daily assignments topractice at home (Daily Exercises and Hierarchy Exercises);treatment days (one other time for 5–10 minutes);nontreatment days (two times for 10–15 minutes);homework book provided and patient made accountable

Treatment: focus attention on how it feels and looks to moveBIGCarryover activities: start day one; daily assignments(treatment and nontreatment days); use big movements inreal-life situations; (i) difficulty of the assignment matchesthe level of the hierarchy where the person is working; (ii)make patient accountable and probe for comments frompatient that people in their daily living have said, such as,“You are moving better”Homework practice: start day one: daily assignments topractice at home (Daily Exercises and FunctionalComponent Movements, Walking BIG); treatment days (oneother time for 5–10 minutes); nontreatment days (two timesfor 10–15 minutes); homework book provided and patientmade accountable

these hypotheses. These initial neural findings underlyingLSVT LOUD outcomes are being further examined andverified in ongoing imaging studies as discussed in ongoingresearch.

4. What Is LSVT BIG?

Individuals with PD perform movements that are hesitant(akinesia), slow (bradykinesia), and with reduced ampli-tude (hypokinesia). Changing from one motor programto another (set-shifting) may be disturbed and sequencingof repetitive movements may occur with prolonged and/orirregular intervals and reduced and/or irregular amplitudes[56]. External cues may exert disproportionate influenceson motor performance and can trigger both motor blocksand kinesia paradoxica [57]. In LSVT BIG, training ofamplitude rather than speed was chosen as the mainfocus of treatment to overcome bradykinesia/hypokinesiabecause training of velocity can induce faster movementsbut does not consistently improve movement amplitude andaccuracy. Furthermore, training to increase velocity of limbmovements may result in hypokinetic (reduced) movementamplitude [58, 59]. In contrast, training of amplitude notonly results in bigger, but also in faster and more precisemovement [58, 59]. The goal of LSVT BIG is to overcomedeficient speed-amplitude regulation leading to underscalingof movement amplitude at any given velocity [59–61].Continuous feedback on motor performance and trainingof movement perception is used to counteract reduced gainin motor activities resulting from disturbed sensorimotorprocessing [62].

Most current therapies rely on compensatory behaviorand external cueing in order to bypass deficient basal gangliafunction [58, 63–70]. In contrast, other protocols focus onretraining of deficient functions. Task-specific, repetitive,high-intensity exercises for individuals with PD includetreadmill training [71], training of compensatory steps [72]walking [73], and muscle strengthening [74, 75]. LSVT BIGbelongs to the latter restorative approaches and is aiming

to restore normal movement amplitude by recalibratingthe patient’s perception of movement execution. LSVT BIGdiffers from other forms of physiotherapy in PD in its train-ing of movement amplitude as a single treatment parameter(both single motor target and cognitive cue) through higheffort, intensive treatment with a focus on recalibratingsensory perception of normal amplitude of movements. Thestandardized protocol of LSVT BIG was derived directly fromLSVT LOUD and is summarized in Table 1.

5. LSVT BIG Outcome Data

Presently two trials on the effectiveness of LSVT BIG havebeen published.

A noncontrolled study assessed effects of LSVT BIG in18 individuals with PD [76]. Data documented that afterfour weeks of training, subjects demonstrated a modest(12%–14%) increase in velocity of walking and reachingmovements.

In the recently published rater-blinded Berlin LSVT BIGStudy improvement in motor performance was comparedin 60 individuals with PD, randomly assigned to receiveLSVT BIG, Nordic Walking (as group treatment) or domestictraining without supervision [60]. Mean improvement ofUPDRS motor score in subjects receiving LSVT BIG was 5.05at four-month followup. In contrast, the UPDRS motor scoreslightly deteriorated in control groups undergoing training inNordic walking with the same amount of supervised sessionsand in subjects who received domestic training receiving a1-hour instructional lesson and no further supervision bya therapist. The beneficial outcome in LSVT BIG was alsoreflected by improvements in further assessments including astandard time-up and go task and 10-meter walk. Accordingto Schrag et al. [77] a change of five points is the mostappropriate cutoff score for the minimal clinical importantchange (MCIC) of the UPDRS motor score for all Hoehnand Yahr stages from Stage I to III. The degree of change inUPDRS motor score after LSVT BIG can thus be assumedto be clinically relevant. There is no established definition


of the MCIC for the secondary motor assessments, but theobserved 10–15% improvements in timed test of mobility arelikely to have functional impact.

The Berlin LSVT BIG Study is one of the few studiescomparing specific types of physiotherapy with both activecomparators and inactive controls. Sage and Almeida [78]reported more improvement in the UPDRS motor scoreand other motor tasks with exercises designed to improvesensory attention and body awareness when compared tolower-limb aerobic training. Mak and Hui-Chan 2008 [79]found better outcomes in the Sit-and-Stand task whensubjects received training including sensory as compared toconventional exercise. In both studies individuals withoutactive interventions did not improve. In the Berlin LSVT BIGStudy outcomes differed clearly between active interventions.Intensive one-to-one training (LSVT BIG) was found tobe more effective than Nordic walking delivered as grouptraining. Although differences in training techniques mayalso have influenced results, it is likely that the specificprotocol of LSVT BIG and, possibly, individual face-to-faceinteraction between patient and therapist, was more crucialfor successful outcomes than total exercise time. Furtherstudies are needed to explore differences in cost-effectivenessbetween the more expensive individual LSVT BIG training,group treatments, and self-supervised domestic exercise.

6. Unique Fundamentals of LSVT Programs

6.1. Target: Amplitude. We hypothesize that training-induced increases in movement amplitude target the pro-posed pathophysiological mechanisms underlying bradykin-esia/hypokinesia—inadequate muscle activation [62]. Themuscle activation deficits that occur in bradykinesia arebelieved to result from inadequate merging of kinestheticfeedback, motor output, and context feedback within thebasal ganglia, necessary to select and reinforce an appropriategain in the motor command [62, 80]. Although the targetis increased amplitude, it is important to note that theend result in speech and movement amplitude output(louder voice/bigger movements) is within normal limits.The cue of “loud” or “big” is used to simply drive increasedmotor output across the motor systems for more normalamplitude. The role of the speech, physical, or occupationaltherapist is to shape the amplitude into healthy, goodquality movements (see Shaping in Table 1). Post-LSVTLOUD videostroboscopic data [47] and perceptual ratingsof voice [33] indicate improved laryngeal function and voicequality rather than vocal hyperfunction or deterioration invoice posttreatment. Ratings of motor performance afterLSVT BIG also indicated a trend towards normality and noexaggeration or overcompensation of movement amplitudes[60, 76, 81].

The idea of targeting amplitude in rehabilitation forindividuals with PD is not new. Training vocal loudness(amplitude) is consistent with approaches recommended fortreating motor speech disorders that (a) create a single motororganizing theme, (b) have a maximum impact on otheraspects of speech production, and (c) increase effort acrossthe speech mechanism [81–83]. Further, many physical

therapy programs have amplitude as a component of therapyeither as exercise principles or by using external cues (e.g.,[84, 85]). The unique element of training amplitude in LSVTPrograms is that it is the exclusive focus. We hypothesize thata single, overlearned cue (louder voice/bigger movements)may minimize cognitive load and mental effort [86] and pos-sibly facilitate maintenance and generalization of treatmentstrategies outside of the therapy room. This hypothesis isyet to be formally tested and is an area for future research.For example, testing the impact of dual task functioningon the ability of individuals with PD to maintain improvedamplitude before/after LSVT Programs would elucidate theability of these individuals to learn a new self-cue foramplitude.

6.2. Mode: Intensive, High Effort Therapy. The trainingmode of LSVT Programs is consistent with some principlesthat promote activity-dependent neuroplasticity [11, 87]including (a) specificity, targeting bradykinesia/hypokinesiathrough increasing amplitude of motor output, (b) intensity,increased dosage of treatment, (c) repetition, increased repe-tition of tasks (minimum 15 repetitions) within treatmentsessions and home practice, and (d) saliency of treatmenttasks, individualized hierarchy and carryover assignmentsfor active practice of desired goals, interests and abilities ofeach person [9–16]. Further, we recognize that acquisitionof the motor skill (e.g., louder voice, bigger movements)alone may not be sufficient for sustained neuroplasticity(i.e., sensorimotor map reorganization, synaptogenesis) [14]or for carryover and long-term maintenance outside thetherapeutic environment. Therefore, a direct translationof the structured motor exercises (daily exercises) intofunctional daily activities is emphasized in treatment withthe goal of facilitating generalization outside of the treatmentroom (see Table 1 Hierarchy, Carryover and Homework).In addition, emphasis is placed on establishing life-longhabits of structured homework practice of voice/movementexercises that continue beyond the one-month of treatment.Finally, simply using the louder voice or bigger movementsin daily living provides additional practice, as summarizedby this patient quote,

“in my normal everyday life, I just exaggerate mymovements. I keep things big when I reach forthings, or when I bend or when I walk; and whenI talk–I keep my voice loud.”

6.3. Recalibration: Addressing Barriers to Generalization. Sen-sorimotor processing deficits during speech and movementhave been well documented [37, 38, 88–91]. From our ownclinical observations, it appears that addressing the motordeficit in isolation is not sufficient for lasting treatmentoutcomes that generalize beyond the treatment room [34].Thus, the LSVT Programs are designed to train individualswith PD to recalibrate their motor and perceptual systems sothat they are less inclined to downscale (reduce amplitude)speech and limb movement parameters after treatment.

Figure 1 illustrates our hypothesized model for ampli-tude rescaling and recalibration in LSVT Programs. In short,


Problem in self-perception/awareness:do not recognize movements

are soft, small, or slow

Self-cueing deficits:continue scaling reduced amplitudeof speech and movement patterns

Produce soft voice,small, slow movements

Reducedamplitude of motor output

Pretreatment

(a)

Produce louder voice, larger movements

Increaseamplitude of motor output

Improve self-perception/awarenessof amplitude required to

produce normal vocal loudness andmovement amplitude

Improve self-cueing/attention to action:habitually scale increased amplitude

of speech and movement patterns

Treatment focus: mode of delivery is intensive, high effort, and salient

(b)

Figure 1: We hypothesize that pretreatment (a), individuals with PD have reduced amplitude of motor output, which results in soft voiceand small movements. Due to problems in sensory self-perception they are not aware of the soft voice and small movements, or they do notrecognize the extent of their soft voice and smaller movements. As a result, no error correction is made and individuals continue to programor self-cue reduced amplitude of motor output. They are “stuck” in a cycle of being soft and small. The focus in treatment (b) is on increasingthe amplitude of motor output by having individuals with PD produce a louder voice and larger movements. Individuals are then taughtthat what feels/sounds/looks “too loud” or “too big” is within normal limits and has a positive impact on daily functional living. Thereforeat the end of treatment, individuals habitually self-cue increased amplitude of motor output and have attention to action. Now they are in acycle of a louder voice and bigger movements.

the goal is to teach individuals with PD to produce motoroutput required for louder voice and bigger movements(Figure 1(b)) and help them recognize that this increasedoutput results in within normal limits voice and movements.Directly addressing this sensory mismatch may help indi-viduals learn to habitually (i.e., self-cue) speak with greatervocal loudness and move with bigger movements at theend of therapy. A specific example of a recalibration taskincludes recording the individual’s voice while reading in avoice that they self-perceive as “too loud” and then playingit back to them. Individuals with PD can recognize whenthey hear the audio recordings that what felt and soundedtoo loud to them while reading, actually sounds withinnormal limits (or in some cases still too soft). Similarly,video recording an individual with PD as they walk or

move in a manner that they perceive as “too big” allowsthem to visualize that what felt too big to them actuallylooks like normal movements (or in some cases still toosmall). Additional recalibration activities are detailed inTable 1.

The hypothesized concepts underlying recalibration inLSVT Programs have yet to be systematically tested inpre/posttreatment experiments. However, there is evidencethat cognitive training is possible in individuals with PD[92], including training in motor attention to action andperformance under multiple tasks [93, 94]. Moreover, theability to speak in a louder voice two years after interventionas compared to pretreatment levels [27] support the abilityof LSVT-LOUD-trained individuals to self-monitor vocalloudness at some level.


7. Limitations of LSVT Programs

There are a number of limitations to the scope of researchon LSVT Programs and we have highlighted some of thekey areas below. First, there is a need to better defineprognostic variables for who will respond best to LSVTPrograms and what outcomes can be expected in individualswith a variety of factors, such as depression, dementia,apathy, orthopedic complications, and dyskinesias, as wellas atypical PD and post-DBS surgery. While the majorityof LSVT outcome data have been reported on individ-uals with idiopathic PD, single subject, case study andsmall group designs have documented post-LSVT-LOUDimprovements in individuals after neurosurgery and withatypical parkinsonism [95–97]. However, these outcomesmay not be of the same magnitude as those observed inindividuals with mild-to-moderate idiopathic PD and theseindividuals may require more frequent follow-up treatmentsessions to maintain improvements over time. Furthermore,LSVT LOUD outcomes in individuals with significant ratedisorders, such as palilalia, and individuals who havesevere speech disorders secondary to high-frequency DBSstimulation have been poor. Data examining LSVT BIG inatypical and post-DBS populations are not available. Second,studies examining the optimal dose-response relationshipsfor LSVT Programs across idiopathic PD, atypical PD, andindividual post-DBS are needed. The standard dose of LSVTPrograms is 16 individual 60-minute sessions within onemonth. There is one dose-response study for LSVT LOUDthat examined the impact of an extended treatment protocol(LSVT Extended, LSVT-X) [98]. Specifically, individualsreceived in-person treatment two days a week and completedhome practice sessions the other two days a week for 8 weeksof treatment. Outcome data immediately posttreatment werecomparable to the standard dosage. Of note, the treatingclinicians completed daily calls and extensive home-practicemonitoring to ensure that all subjects completed all homesessions. Ongoing work is examining alternative dosages ofLSVT BIG, additional dose-response relationships need tobe defined. Third, the spread of effects across the speechproduction system has been reported following LSVT LOUD.These studies should be further advanced and studies areneeded to evaluate the spread of effects or transfer effectsfrom large body movements to fine motor functions, balance,or dual tasks following LSVT BIG. Fourth, the practicaland financial feasibility of delivering intensive treatmentin LSVT Programs must be addressed. Physical immobilityand geographical constraints are barriers which limit patientaccessibility to intensive treatment. Fifth, the maintenanceand enhancement of long-term treatment effects also areareas of need. While LSVT LOUD outcome data reportmaintenance of treatment effects for two years after onemonth of treatment, we believe outcomes can be furtheroptimized. The long-term effects of LSVT BIG need tobe established. Strategies to maximize compliance withcontinued home practice and the timing of optimal follow-up treatment intervals need to be defined. Finally, thehypothesized concepts underlying sensory calibration as wellas understanding neural mechanisms of treatment-related

change need to be systematically studied and validated.Only then can we fully understand what elements of treat-ment contribute to improvement in speech and movementfunctioning. These limitations will continue to guide ourfuture research with some areas already being addressed asdiscussed below.

8. Current and Future Research Directions

Our ongoing work in LSVT LOUD is addressing questionsrelated to the importance of the treatment target versusthe mode of delivery. Specifically, we are comparing twotreatment targets: vocal loudness training (LSVT LOUD)versus orofacial/articulation training (LSVT ARTIC) andthe effects on measures of speech intelligibility, speechacoustics, facial expression, and swallowing. The two treat-ments are standardized and matched in terms of modeof delivery (e.g., dosage, sensory recalibration, homework,and carryover assignments). LSVT LOUD focuses on train-ing healthy vocal loudness across speech tasks (sustainedvowels, high/low vowels, functional phrases, and speechhierarchy), with focused attention on how it feels andsounds to talk LOUD, whereas, LSVT ARTIC focuses onhigh-force articulation or enunciation across speech tasks(diadochokinesis, contrastive pairs, functional phrases, andspeech hierarchy), with focus on how it feels to have high-effort enunciation. Preliminary data examining single-wordintelligibility in noise conditions [99] and facial expressionsutilizing the Facial Action Coding System (FACS) [100]revealed significant improvements from pre to posttreatmentin the LSVT LOUD group only. More extensive analysisis ongoing. In addition, this study includes comprehensiveneuropsychological profiles of subjects and may shed somelight on the impact of factors such as age, stage of disease,depression, dementia, or other nonmotor symptoms ontreatment outcomes.

The impact of DBS on speech is an urgent area ofresearch. Tripoliti and colleagues [101] are assessing thereasons for the heterogeneous speech outcomes followingDBS-STN by involving simultaneous quantitative measuresof pre- and postsurgical speech functioning and detailsof surgical and stimulator optimization. Knowledge gainedfrom these studies is likely to facilitate development oftreatment approaches for speech problems in individualswith DBS-STN either before surgery (as preventative) or aftersurgery (as rehabilitation). Our laboratory is looking at theimpact of additional weeks of treatment on speech outcomesfor individuals with PD after DBS.

Advances in computer and web-based technology offerpotentially powerful solutions to the problems of treat-ment accessibility, efficacious dosage delivery, and long-term maintenance in rehabilitation [102–104]. Preliminarystudies have documented the impact of telepractice andsoftware programs on treatment availability for LSVT LOUDand suggest that such technology may be effective [42, 105–108] and increase the feasibility of intensive dosage and long-term followup. In addition, a study by Tindall et al. [105]completed a cost analysis comparing in-person delivery ofLSVT LOUD versus telepractice delivery. The computed


mean amount of time and money for individuals withPD across these two modes of delivery was reported. Thelive delivery mode required 51 hours for 16 visits (traveland therapy time), $953.00 on fuel/mileage expenses, and$269.00 for other expenses (e.g., food). In contrast, thetelepractice delivery option required 16 hours of time(therapy, no travel) and no additional costs for fuel/mileageor other expenses. To further enhance accessibility, a softwareprogram designed to collect acoustic data and provideinteractive feedback as it guides the patient through theLSVT LOUD exercises has been developed. Outcome datadocument that treatment effects are comparable when half ofthe sessions were delivered by software [109]. These studiesneed further validation. While telepractice has not beenexplored for delivery of LSVT BIG, there are studies thathave documented the feasibility of remote measuring ofactivities of daily living [110] and ongoing trials examiningthe delivery of physical therapy via telepractice in patientsafter stroke [111]. Thus, future applications of both teleprac-tice and software programs/gaming technology to increaseaccessibility and feasibility of LSVT BIG is possible. The useof technology is not LSVT specific and may have the abilityto increase accessibility, enhance effectiveness, and reducefinancial burden of many intensive rehabilitation programsfor people with PD.

Understanding neural mechanisms of both speech andmovement disorders in PD as well as mechanism oftreatment-related change are of great promise to helpimprove treatment outcomes. As part of our ongoingwork we are examining neural changes (PET imaging)in individuals with PD across the LSVT LOUD, LSVTARTIC, and Untreated groups. Hypothetically, intensivepractice of speech enunciation by the LSVT ARTIC regimenshould strengthen cortically mediated speech articulationin PD, beyond the improvement associated with LSVTLOUD. To our knowledge, this will be the first imagingstudy of comparison speech treatments in individuals withPD including long-term followup (3 months). Developingparallel imaging studies before/after LSVT BIG is of greatinterest to us both in terms of understanding reorganizationof brain activation patterns following treatment but alsoto understand differences between using amplitude to treatspeech versus limb motor systems.

Finally, whereas studies of movement and limb/gaitexercise in animal models of PD have contributed immenselyto the literature, there have been no analogous models forstudying vocalization deficits. Today, emerging models ofvocal motor deficits following dopamine depletion in rodents(rats and mice) and songbirds offer promise for the feasibilityand value of these models [112–114]. Viable animal modelsof vocalization patterns associated with PD may allow us toaccelerate the acquisition of the neurobiological and behav-ioral evidence to improve our understanding of voice/speechdeficits in PD and document the therapeutic value of earlyinterventions to slow voice/speech symptom progression inhuman PD.

Collectively these ongoing studies have the potential toimprove our understanding of the underlying mechanismsof speech-treatment-related changes in individuals with

PD and will help guide treatment improvements. Futureresearch will address the underlying bases for treatment-related changes that have a beneficial impact on speechand movement and thus quality of life in individuals withParkinson disease.

Disclosure

L. Ramig and C. Fox receive lecture honoraria and haveownership interest in LSVT Global, Inc. They are in fullcompliance with Federal Statute (42 C.F.R. Part 50. SubpartF) and the University of Colorado-Boulder Policy on Conflictof Interest and Commitment.

Acknowledgments

The work described here was funded in part by Grant R01DC-01150 from the National Institutes of Health, NationalInstitute of Deafness and other Communication Disorders(NIH-NIDCD).

References

[1] S. H. J. Keus, M. Munneke, M. J. Nijkrake, G. Kwakkel, and B.R. Bloem, “Physical therapy in Parkinson’s disease: evolutionand future challenges,” Movement Disorders, vol. 24, no. 1,pp. 1–14, 2009.

[2] N. Miller, E. Noble, D. Jones, and D. Burn, “Life withcommunication changes in Parkinson’s disease,” Age andAgeing, vol. 35, no. 3, pp. 235–239, 2006.

[3] G. Ransmayr, “Physical, occupational, speech and swallowingtherapies and physical exercise in Parkinson’s disease,” Jour-nal of Neural Transmission, vol. 118, no. 5, pp. 773–781, 2011.

[4] B. E. Fisher, A. D. Wu, G. J. Salem et al., “The effect of exercisetraining in improving motor performance and corticomotorexcitability in people with early Parkinson’s disease,” Archivesof Physical Medicine and Rehabilitation, vol. 89, no. 7, pp.1221–1229, 2008.

[5] M. A. Hirsch and B. G. Farley, “Exercise and neuroplasticityin persons living with Parkinson’s disease,” European Journalof Physical and Rehabilitation Medicine, vol. 45, no. 2, pp.215–229, 2009.

[6] G. M. Petzinger, B. E. Fisher, J. E. Van Leeuwen et al.,“Enhancing neuroplasticity in the basal ganglia: the role ofexercise in Parkinson’s disease,” Movement Disorders, vol. 25,supplement 1, pp. S141–S145, 2010.

[7] J. L. Tillerson and G. W. Miller, “Forced limb-use andrecovery following brain injury,” Neuroscientist, vol. 8, no. 6,pp. 574–585, 2002.

[8] J. L. Tillerson, W. M. Caudle, M. E. Reveron, and G. W.Miller, “Exercise induces behavioral recovery and attenuatesneurochemical deficits in rodent models of Parkinson’sdisease,” Neuroscience, vol. 119, no. 3, pp. 899–911, 2003.

[9] J. L. Tillerson, A. D. Cohen, J. Philhower, G. W. Miller, M. J.Zigmond, and T. Schallert, “Forced limb-use effects on thebehavioral and neurochemical effects of 6-hydroxydopam-ine,” Journal of Neuroscience, vol. 21, no. 12, pp. 4427–4435,2001.

[10] J. A. Kleim and T. A. Jones, “Principles of experience-dependent neural plasticity: implications for rehabilitationafter brain damage,” Journal of Speech, Language, and HearingResearch, vol. 51, no. 1, pp. S225–S239, 2008.


[11] J. A. Kleim, T. A. Jones, and T. Schallert, “Motor enrichmentand the induction of plasticity before or after brain injury,”Neurochemical Research, vol. 28, no. 11, pp. 1757–1769, 2003.

[12] M. G. Vucckovic, Q. Li, B. Fisher et al., “Exercise elevatesdopamine D2 receptor in a mouse model of Parkinson’sdisease: in vivo imaging with [18F]fallypride,” MovementDisorders, vol. 25, no. 16, pp. 2777–2784, 2010.

[13] J. M. Conner, A. A. Chiba, and M. H. Tuszynski, “The basalforebrain cholinergic system is essential for cortical plasticityand functional recovery following brain injury,” Neuron, vol.46, no. 2, pp. 173–179, 2005.

[14] J. A. Kleim, T. M. Hogg, P. M. VandenBerg, N. R. Cooper,R. Bruneau, and M. Remple, “Cortical synaptogenesis andmotor map reorganization occur during late, but not early,phase of motor skill learning,” Journal of Neuroscience, vol.24, no. 3, pp. 628–633, 2004.

[15] J. A. Kleim, N. R. Cooper, and P. M. VandenBerg, “Exerciseinduces angiogenesis but does not alter movement represen-tations within rat motor cortex,” Brain Research, vol. 934, no.1, pp. 1–6, 2002.

[16] M. S. Remple, R. M. Bruneau, P. M. VandenBerg, C.Goertzen, and J. A. Kleim, “Sensitivity of cortical movementrepresentations to motor experience: evidence that skilllearning but not strength training induces cortical reorgani-zation,” Behavioural Brain Research, vol. 123, no. 2, pp. 133–141, 2001.

[17] M. J. Zigmond, J. L. Cameron, R. K. Leak et al., “Triggeringendogenous neuroprotective processes through exercise inmodels of dopamine deficiency,” Parkinsonism and RelatedDisorders, vol. 15, no. 3, pp. S42–S45, 2009.

[18] A. K. Ho, R. Iansek, C. Marigliani, J. L. Bradshaw, and S.Gates, “Speech impairment in a large sample of patients withParkinson’s disease,” Behavioural Neurology, vol. 11, no. 3, pp.131–137, 1998.

[19] J. A. Logemann, H. B. Fisher, B. Boshes, and E. R. Blonsky,“Frequency and cooccurrence of vocal tract dysfunctions inthe speech of a large sample of Parkinson patients,” Journal ofSpeech and Hearing Disorders, vol. 43, no. 1, pp. 47–57, 1978.

[20] F. L. Darley, A. E. Aronson, and J. R. Brown, “Clusters ofdeviant speech dimensions in the dysarthrias,” Journal ofSpeech and Hearing Research, vol. 12, no. 3, pp. 462–496,1969.

[21] S. Sapir, A. A. Pawlas, L. O. Ramig et al., “Voice and speechabnormalities in Parkinson disease: relation to severity ofmotor impairment, duration of disease, medication, depres-sion, gender, and age,” Journal of Medical Speech-LanguagePathology, vol. 9, no. 4, pp. 213–226, 2001.

[22] J. Muller, G. K. Wenning, M. Verny et al., “Progression ofdysarthria and dysphagia in postmortem-confirmed parkin-sonian disorders,” Archives of Neurology, vol. 58, no. 2, pp.259–264, 2001.

[23] J. Rusz, R. Cmejla, H. Ruzickova, and E. Ruzicka, “Quanti-tative acoustic measurements for characterization of speechand voice disorders in early untreated Parkinson’s disease,”Journal of the Acoustical Society of America, vol. 129, no. 1,pp. 350–367, 2011.

[24] M. T. Sarno, “Speech impairment in Parkinson’s disease,”Archives of Physical Medicine and Rehabilitation, vol. 49, no.5, pp. 269–275, 1968.

[25] L. Ramig, A. Pawlas, and S. Countryman, Lee Silverman VoiceTreatment: A Practical Guide to Treating the Voice and SpeechDisorders in Parkinson Disease, National Center for Voice andSpeech, Iowa City, IA, USA, 1995.

[26] L. O. Ramig and C. Dromey, “Aerodynamic mechanismsunderlying treatment-related changes in vocal intensity inpatients with Parkinson disease,” Journal of Speech, Language,and Hearing Research, vol. 39, no. 4, pp. 798–807, 1996.

[27] L. O. Ramig, S. Sapir, S. Countryman et al., “Intensive voicetreatment (LSVT�) for individuals with Parkinson’s disease:A two year follow-up,” Journal of Neurology Neurosurgery andPsychiatry, vol. 71, no. 4, pp. 493–498, 2001.

[28] L. O. Ramig, S. Sapir, C. Fox, and S. Countryman,“Changes in vocal loudness following intensive voice treat-ment (LSVT�) in individuals with Parkinson’s disease: acomparison with untreated patients and nornal age-matchedcontrols,” Movement Disorders, vol. 16, no. 1, pp. 79–83,2001.

[29] I. Titze, Vocal Fold Physiology: Frontiers in Basic Science,Singular, San Diego, Calif, USA, 1993.

[30] L. O. Ramig, S. Countryman, L. L. Thompson, and Y. Horii,“Comparison of two forms of intensive speech treatment forParkinson disease,” Journal of Speech and Hearing Research,vol. 38, no. 6, pp. 1232–1251, 1995.

[31] S. Sapir, J. L. Spielman, L. O. Ramig, B. H. Story, and C.Fox, “Effects of intensive voice treatment (the Lee SilvermanVoice Treatment [LSVT]) on vowel articulation in dysarthricindividuals with idiopathic Parkinson disease: acoustic andperceptual findings,” Journal of Speech, Language, and Hear-ing Research, vol. 50, no. 4, pp. 899–912, 2007.

[32] S. Sapir, L. O. Ramig, J. L. Spielman, and C. Fox, “Formantcentralization ratio: a proposal for a new acoustic measure ofdysarthric speech,” Journal of Speech, Language, and HearingResearch, vol. 53, no. 1, pp. 114–125, 2010.

[33] C. A. Baumgartner, S. Sapir, and L. O. Ramig, “Voice qualitychanges following phonatory-respiratory effort treatment(LSVT�) versus respiratory effort treatment for individualswith Parkinson disease,” Journal of Voice, vol. 15, no. 1, pp.105–114, 2001.

[34] C. M. Fox, C. E. Morrison, L. O. Ramig, and S. Sapir,“Current perspectives on the Lee Silverman voice treatment(LSVT�) for individuals with idiopathic Parkinson disease,”American Journal of Speech-Language Pathology, vol. 11, no.2, pp. 111–123, 2002.

[35] G. E. Alexander and M. D. Crutcher, “Functional architectureof basal ganglia circuits: neural substrates of parallel pro-cessing,” Trends in Neurosciences, vol. 13, no. 7, pp. 266–271,1990.

[36] A. M. Graybiel, “The basal ganglia: learning new tricks andloving it,” Current Opinion in Neurobiology, vol. 15, no. 6, pp.638–644, 2005.

[37] A. K. Ho, J. L. Bradshaw, and R. Iansek, “Volume perceptionin Parkinsonian speech,” Movement Disorders, vol. 15, no. 6,pp. 1125–1131, 2000.

[38] A. K. Ho, J. L. Bradshaw, R. Iansek, and R. Alfredson, “Speechvolume regulation in Parkinson’s disease: effects of implicitcues and explicit instructions,” Neuropsychologia, vol. 37, no.13, pp. 1453–1460, 1999.

[39] B. G. Farley, C. M. Fox, L. O. Ramig, and D. H. McFar-land, “Intensive amplitude-specific therapeutic approachesfor Parkinson’s disease: toward a neuroplasticity-principledrehabilitation model,” Topics in Geriatric Rehabilitation, vol.24, no. 2, pp. 99–114, 2008.

[40] S. Sapir, L. O. Ramig, and C. M. Fox, “Intensive voice treat-ment in Parkinson’s disease: Lee Silverman voice treatment,”Expert Review of Neurotherapeutics, vol. 11, no. 6, pp. 815–830, 2011.


[41] G. Constantinescu, D. Theodoros, T. Russell, E. Ward, S.Wilson, and R. Wootton, “Treating disordered speech andvoice in Parkinson’s disease online: a randomized controllednon-inferiority trial,” International Journal of Language andCommunication Disorders, vol. 46, no. 1, pp. 1–16, 2011.

[42] S. Howell, E. Tripoliti, and T. Pring, “Delivering the Lee Sil-verman Voice Treatment (LSVT) by web camera a feasibilitystudy,” International Journal of Language and CommunicationDisorders, vol. 44, no. 3, pp. 287–300, 2009.

[43] D. G. Theodoros, E. C. Thompson-Ward, B. E. Murdoch, J.Lethlean, and P. Silburn, “The effects of the Lee SilvermanVoice Treatment Program on motor speech function inParkinson disease following thalamotomy and pallidotomysurgery: a case study,” Journal of Medical Speech-LanguagePathology, vol. 7, no. 2, pp. 157–160, 1999.

[44] E. C. Ward, D. G. Theodoros, B. E. Murdoch, and P. Silburn,“Changes in maximum capacity tongue function followingthe Lee Silverman Voice Treatment Program,” Journal ofMedical Speech-Language Pathology, vol. 8, no. 4, pp. 331–335, 2000.

[45] K. M. Yorkston, M. Hakel, D. R. Beukelman, and S. Fager,“Evidence for effectiveness of treatment of loudness, rate, orprosody in dysarthria: a systematic review,” Journal of MedicalSpeech-Language Pathology, vol. 15, no. 2, pp. xi–xxxvi, 2007.

[46] J. E. Huber, E. T. Stathopoulos, L. O. Ramig, and S. L. Lan-caster, “Respiratory function and variability in individualswith parkinson disease: pre- and post-Lee Silverman voicetreatment,” Journal of Medical Speech-Language Pathology,vol. 11, no. 4, pp. 185–201, 2003.

[47] M. E. Smith, L. O. Ramig, C. Dromey, K. S. Perez, and R.Samandari, “Intensive voice treatment in Parkinson disease:laryngostroboscopic findings,” Journal of Voice, vol. 9, no. 4,pp. 453–459, 1995.

[48] C. Dromey, L. O. Ramig, and A. B. Johnson, “Phonatory andarticulatory changes associated with increased vocal intensityin Parkinson disease: a case study,” Journal of Speech andHearing Research, vol. 38, no. 4, pp. 751–763, 1995.

[49] J. L. Spielman, J. C. Borod, and L. O. Ramig, “The effects ofintensive voice treatment on facial expressiveness in Parkin-son disease: preliminary data,” Cognitive and BehavioralNeurology, vol. 16, no. 3, pp. 177–188, 2003.

[50] A. El Sharkawi, L. Ramig, J. A. Logemann et al., “Swallowingand voice effects of Lee Silverman Voice Treatment: a pilotstudy,” Journal of Neurology Neurosurgery and Psychiatry, vol.72, no. 1, pp. 31–36, 2002.

[51] K. Forrest, G. Weismer, and G. S. Turner, “Kinematic, acous-tic, and perceptual analyses of connected speech producedby Parkinsonian and normal geriatric adults,” Journal of theAcoustical Society of America, vol. 85, no. 6, pp. 2608–2622,1989.

[52] M. D. McClean and S. M. Tasko, “Association of orofacialwith laryngeal and respiratory motor output during speech,”Experimental Brain Research, vol. 146, no. 4, pp. 481–489,2002.

[53] M. Liotti, D. Vogel, S. Sapir, L. Ramig, P. New, and P.Fox, “Abnormal auditory gating in Parkinson’s disease before& after LSVT�,” in Proceedings of the Annual Meeting ofthe American Speech, Language and Hearing Association,Washington, DC, USA, 2000.

[54] M. Liotti, L. O. Ramig, D. Vogel et al., “Hypophonia inParkinson’s disease: neural correlates of voice treatmentrevealed by PET,” Neurology, vol. 60, no. 3, pp. 432–440, 2003.

[55] S. Narayana, P. T. Fox, W. Zhang et al., “Neural correlates ofefficacy of voice therapy in Parkinson’s disease identified by

performance-correlation analysis,” Human Brain Mapping,vol. 31, no. 2, pp. 222–236, 2010.

[56] N. Georgiou, R. Iansek, J. L. Bradshaw, J. G. Phillips, J.B. Mattingley, and J. A. Bradshaw, “An evaluation of therole of internal cues in the pathogenesis of Parkinsonianhypokinesia,” Brain, vol. 116, no. 6, pp. 1575–1587, 1993.

[57] B. R. Bloem, J. M. Hausdorff, J. E. Visser, and N. Giladi, “Fallsand freezing of Gait in Parkinson’s disease: a review of twointerconnected, episodic phenomena,” Movement Disorders,vol. 19, no. 8, pp. 871–884, 2004.

[58] A. L. Behrman, P. Teitelbaum, and J. H. Cauraugh, “Verbalinstructional sets to normalise the temporal and spatialgait variables in Parkinson’s disease,” Journal of NeurologyNeurosurgery and Psychiatry, vol. 65, no. 4, pp. 580–582,1998.

[59] M. E. Morris, R. Iansek, T. A. Matyas, and J. J. Summers,“The pathogenesis of gait hypokinesia in Parkinson’s disease,”Brain, vol. 117, no. 5, pp. 1169–1181, 1994.

[60] G. Ebersbach, M. Sojer, F. Valldeoriola et al., “Comparativeanalysis of gait in Parkinson’s disease, cerebellar ataxia andsubcortical arteriosclerotic encephalopathy,” Brain, vol. 122,no. 7, pp. 1349–1355, 1999.

[61] F. B. Horak, J. Frank, and J. Nutt, “Effects of dopamine onpostural control in parkinsonian subjects: scaling, set, andtone,” Journal of Neurophysiology, vol. 75, no. 6, pp. 2380–2396, 1996.

[62] A. Berardelli, J. C. Rothwell, P. D. Thompson, and M. Hallett,“Pathophysiology of bradykinesia in parkinson’s disease,”Brain, vol. 124, no. 11, pp. 2131–2146, 2001.

[63] R. Marchese, M. Diverio, F. Zucchi, C. Lentino, and G.Abbruzzese, “The role of sensory cues in the rehabilitation ofParkinsonian patients: a comparison of two physical therapyprotocols,” Movement Disorders, vol. 15, no. 5, pp. 879–883,2000.

[64] A. Nieuwboer, G. Kwakkel, L. Rochester et al., “Cueingtraining in the home improves gait-related mobility inParkinson’s disease: the RESCUE trial,” Journal of Neurology,Neurosurgery and Psychiatry, vol. 78, no. 2, pp. 134–140,2007.

[65] T. C. Rubinstein, N. Giladi, and J. M. Hausdorff, “Thepower of cueing to circumvent dopamine deficits: a review ofphysical therapy treatment of gait disturbances in Parkinson’sdisease,” Movement Disorders, vol. 17, no. 6, pp. 1148–1160,2002.

[66] M. Dam, P. Tonin, S. Casson et al., “Effects of conventionaland sensory-enhanced physiotherapy on disability of Parkin-son’s disease patients,” Advances in neurology, vol. 69, pp.551–555, 1996.

[67] G. C. McIntosh, S. H. Brown, R. R. Rice, and M. H. Thaut,“Rhythmic auditory-motor facilitation of gait patterns inpatients with Parkinson’s disease,” Journal of NeurologyNeurosurgery and Psychiatry, vol. 62, no. 1, pp. 22–26, 1997.

[68] M. E. Morris, R. Iansek, T. A. Matyas, and J. J. Summers,“Stride length regulation in Parkinson’s disease: normaliza-tion strategies and underlying mechanisms,” Brain, vol. 119,no. 2, pp. 551–568, 1996.

[69] M. H. Thaut, G. C. McIntosh, R. R. Rice, R. A. Miller, J.Rathbun, and J. M. Brault, “Rhythmic auditory stimulationin gait training for Parkinson’s disease patients,” MovementDisorders, vol. 11, no. 2, pp. 193–200, 1996.

[70] V. Muller, B. Mohr, R. Rosin, F. Pulvermuller, F. Muller, andN. Birbaumer, “Short-term effects of behavioral treatmenton movement initiation and postural control in Parkinson’s


disease: a controlled clinical study,” Movement Disorders, vol.12, no. 3, pp. 306–314, 1997.

[71] I. Miyai, Y. Fujimoto, H. Yamamoto et al., “Long-term effectof body weight-supported treadmill training in Parkinson’sdisease: a randomized controlled trial,” Archives of PhysicalMedicine and Rehabilitation, vol. 83, no. 10, pp. 1370–1373,2002.

[72] E. M. Jobges, H. Spittler-Schneiders, C. I. E. Renner,and H. Hummelsheim, “Clinical relevance of rehabilitationprograms for patients with idiopathic Parkinson syndrome.II: symptom-specific therapeutic approaches,” Parkinsonismand Related Disorders, vol. 13, no. 4, pp. 203–213, 2007.

[73] D. A. Lehman, T. Toole, D. Lofald, and M. A. Hirsch,“Training with verbal instructional cues results in near-term improvement of gait in people with Parkinson disease,”Journal of Neurologic Physical Therapy, vol. 29, no. 1, pp. 2–8,2005.

[74] M. A. Hirsch, T. Toole, C. G. Maitland, and R. A. Rider,“The effects of balance training and high-intensity resistancetraining on persons with idiopathic Parkinson’s disease,”Archives of Physical Medicine and Rehabilitation, vol. 84, no.8, pp. 1109–1117, 2003.

[75] L. E. Dibble, T. F. Hale, R. L. Marcus, J. Droge, J. P. Gerber,and P. C. LaStayo, “High-intensity resistance training ampli-fies muscle hypertrophy and functional gains in persons withparkinson’s disease,” Movement Disorders, vol. 21, no. 9, pp.1444–1452, 2006.

[76] B. G. Farley and G. F. Koshland, “Training BIG to movefaster: the application of the speed-amplitude relation as arehabilitation strategy for people with Parkinson’s disease,”Experimental Brain Research, vol. 167, no. 3, pp. 462–467,2005.

[77] A. Schrag, R. Dodel, A. Spottke, B. Bornschein, U. Siebert,and N. P. Quinn, “Rate of clinical progression in Parkinson’sdisease. A prospective study,” Movement Disorders, vol. 22,no. 7, pp. 938–945, 2007.

[78] M. D. Sage and Q. J. Almeida, “Symptom and gait changesafter sensory attention focused exercise vs aerobic training inParkinson’s disease,” Movement Disorders, vol. 24, no. 8, pp.1132–1138, 2009.

[79] M. K. Y. Mak and C. W. Y. Hui-Chan, “Cued task-specifictraining is better than exercise in improving sit-to-stand inpatients with Parkinson’s disease: a randomized controlledtrial,” Movement Disorders, vol. 23, no. 4, pp. 501–509, 2008.

[80] M. Desmurget, S. T. Grafton, P. Vindras, H. Grea, and R. S.Turner, “The basal ganglia network mediates the planning ofmovement amplitude,” European Journal of Neuroscience, vol.19, no. 10, pp. 2871–2880, 2004.

[81] J. Duffy, Motor Speech Disorders, Thieme Medical, New York,NY, USA, 1995.

[82] J. C. Rosenbek and L. L. LaPointe, “The dysarthrias: descrip-tion, diagnosis, and treatment,” in Clinical Management ofNeurogenic Communicative Disorders, D. F. Johns, Ed., pp.97–152, Little, Brown & Co., Boston, Mass, USA, 2nd edition,1985.

[83] K. M. Yorkston, D. R. Beukelman, and K. R. Bell, ClinicalMeasurement of Dysarthric Speakers, College-Hill, Boston,Mass, USA, 1988.

[84] L. A. King and F. B. Horak, “Delaying mobility disability inpeople with parkinson disease using a sensorimotor agilityexercise program,” Physical Therapy, vol. 89, no. 4, pp. 384–393, 2009.

[85] L. Rochester, V. Hetherington, D. Jones et al., “The effectof external rhythmic cues (auditory and visual) on walking

during a functional task in homes of people with Parkinson’sdisease,” Archives of Physical Medicine and Rehabilitation, vol.86, no. 5, pp. 999–1006, 2005.

[86] M. E. Morris, C. L. Martin, and M. L. Schenkman, “Stridingout with Parkinson disease: evidence-based physical therapyfor gait disorders,” Physical Therapy, vol. 90, no. 2, pp. 280–288, 2010.

[87] J. Kleim and T. Jones, “Principles of experience-dependentneural plasticity: implications for rehabilitation after braindamage,” in Proceedings of the Neurorehabilitation WorkshopInnovation and Evolution Workshop, Gainsville, Fla, USA,April 2005.

[88] M. Demirci, S. Grill, L. McShane, and M. Hallett, “Amismatch between kinesthetic and visual perception inParkinson’s disease,” Annals of Neurology, vol. 41, no. 6, pp.781–788, 1997.

[89] E. E. Jobst, M. E. Melnick, N. N. Byl, G. A. Dowling, and M. J.Aminoff, “Sensory perception in Parkinson disease,” Archivesof Neurology, vol. 54, no. 4, pp. 450–454, 1997.

[90] T. Klockgether, M. Borutta, H. Rapp, S. Spieker, and J.Dichgans, “A defect of kinesthesia in Parkinson’s disease,”Movement Disorders, vol. 10, no. 4, pp. 460–465, 1995.

[91] J. S. Schneider, S. G. Diamond, and C. H. Markham,“Parkinson’s disease: sensory and motor problems in armsand hands,” Neurology, vol. 37, no. 6, pp. 951–956, 1987.

[92] A. P. Parıs, H. G. Saleta, M. de la Cruz Crespo Maraveret al., “Blind randomized controlled study of the efficacyof cognitive training in Parkinson’s disease,” MovementDisorders, vol. 26, no. 7, pp. 1251–1258, 2011.

[93] S. G. Brauer and M. E. Morris, “Can people with Parkinson’sdisease improve dual tasking when walking?” Gait andPosture, vol. 31, no. 2, pp. 229–233, 2010.

[94] G. Yogev-Seligmann, N. Giladi, M. Brozgol, and J.M. Haus-dorff, “A training program to improve gait while dual taskingin patients with parkinson’s disease: a pilot study,” Archives ofPhysical Medical Rehabilitation, vol. 93, no. 1, pp. 176–181,2012.

[95] S. Countryman and L. Ramig, “Effects of intensive voice ther-apy on voice deficits associated with bilateral thalamotomy inParkinson’s disease: a case study,” Journal of Medical Speech-Language Pathology, vol. 1, pp. 233–250, 1993.

[96] S. Countryman, L. Ramig, and A. Pawlas, “Speech andvoice deficits in Parkinsonian plus syndromes: can they betreated?” Journal of Medical Speech-Language Pathology, vol.2, no. 3, pp. 211–225, 1994.

[97] J. Spielman, L. Mahler, A. Halpern, P. Gilley, O. Klepitskaya,and L. Ramig, “Intensive voice treatment (LSVTOLOUD) forParkinson’s disease following deep brain stimulation of thesubthalamic nucleus,” Journal of Communication Disorders,vol. 44, no. 6, pp. 688–700, 2011.

[98] J. Spielman, L. O. Ramig, L. Will, A. Halpern, and J. Petska,“Effects of LSVT extended (LSVT-X) on voice and speechin Parkinson disease,” American Journal of Speech, LanguagePathology, vol. 16, no. 2, pp. 95–107, 2007.

[99] A. Halpern, J. Spielman, L. Ramig, I. Panzer, A. Sharpley, andH. Gustafson, “A novel way to measure speech intelligibilityin individuals with Parkinson disease,” Movement Disorders,vol. 26, no. S2, p. S111, 2011.

[100] A. I. Dumer, J. C. Borod, H. Oster, J. L. Spielman, L. A. Rabin,and L. O. Ramig, “Reduction of facial movement deficits inParkinsons disease (PD) after Lee Silverman voice treatment(LSVT). Poster Presentations,” Movement Disorders, vol. 26,pp. S101–S102, 2011.


[101] E. Tripoliti, L. Zrinzo, I. Martinez-Torres et al., “Effectsof contact location and voltage amplitude on speech andmovement in bilateral subthalamic nucleus deep brainstimulation,” Movement Disorders, vol. 23, no. 16, pp. 2377–2383, 2008.

[102] E. Taub, P. S. Lum, P. Hardin, V. W. Mark, and G. Uswatte,“AutoCITE: automated delivery of CI therapy with reducedeffort by therapists,” Stroke, vol. 36, no. 6, pp. 1301–1304,2005.

[103] L. R. Cherney and A. S. Halper, “Novel technology fortreating individuals with aphasia and concomitant cognitivedeficits,” Topics in Stroke Rehabilitation, vol. 15, no. 6, pp.542–554, 2008.

[104] L. M. Manheim, A. S. Halper, and L. Cherney, “Patient-Reported Changes in Communication After Computer-Based Script Training for Aphasia,” Archives of PhysicalMedicine and Rehabilitation, vol. 90, no. 4, pp. 623–627, 2009.

[105] L. R. Tindall, R. A. Huebner, J. C. Stemple, and H. L. Kleinert,“Videophone-delivered voice therapy: a comparative analysisof outcomes to traditional delivery for adults with Parkin-son’s disease,” Telemedicine & E Health, vol. 14, no. 10, pp.1070–1077, 2008.

[106] G. A. Constantinescu, D. G. Theodoros, T. G. Russell, E. C.Ward, S. J. Wilson, and R. Wootton, “Home-based speechtreatment for Parkinson’s disease delivered remotely: a casereport,” Journal of Telemedicine and Telecare, vol. 16, no. 2,pp. 100–104, 2010.

[107] D. G. Theodoros, G. Constantinescu, T. G. Russell, E.C. Ward, S. J. Wilson, and R. Wootton, “Treating thespeech disorder in Parkinson’s disease online,” Journal ofTelemedicine and Telecare, vol. 12, no. 3, pp. 88–91, 2006.

[108] D. Theodoros and L. Ramig, “Telepractice Supported Deliv-ery of LSVT� LOUD,” Perspectives on Neurophysiology andNeurogenic Speech and Language Disorders, vol. 21, no. 3, pp.107–119, 2011.

[109] A. Halpern, C. Matos, L. Ramig, J. Petska, J. Spielman,and L. Will, “Technology supported speech treatment forParkinson’s disease,” Movement Disorders, vol. 20, no. 10, p.S134, 2005.

[110] T. Hoffmann, T. Russell, L. Thompson, A. Vincent, andM. Nelson, “Using the Internet to assess activities of dailyliving and hand function in people with Parkinson’s disease,”NeuroRehabilitation, vol. 23, no. 3, pp. 253–261, 2008.

[111] N. R. Chumbler, D. K. Rose, P. Griffiths et al., “Studyprotocol: home-based telehealth stroke care: a randomizedtrial for veterans,” Trials, vol. 11, article 74, 2010.

[112] M. R. Ciucci, S. T. Ma, C. Fox, J. R. Kane, L. O. Ramig, andT. Schallert, “Qualitative changes in ultrasonic vocalizationin rats after unilateral dopamine depletion or haloperidol: apreliminary study,” Behavioural Brain Research, vol. 182, no.2, pp. 284–289, 2007.

[113] M. R. Ciucci and N. P. Connor, “Dopaminergic influenceon rat tongue function and limb movement initiation,”Experimental Brain Research, vol. 194, no. 4, pp. 587–596,2009.

[114] J. E. Miller, Z. D. Burkett, C. M. Fox, and S. A. White, “Vocalmotor deficits in a songbird model of Parkinson’s disease.Poster Presentations,” Movement Disorders, vol. 26, p. S102,2011.


Review Article

Improving Community Healthcare for Patients withParkinson’s Disease: The Dutch Model

S. H. J. Keus,1 L. B. Oude Nijhuis,2 M.J. Nijkrake,3 B. R. Bloem,2 and M. Munneke1, 4

1 Department of Neurology, Nijmegen Centre for Evidence Based Practice, Radboud University Nijmegen Medical Centre, P.O. Box 9101,6500 HB Nijmegen, The Netherlands

2 Department of Neurology, Donders Institute for Brain, Cognition and Behaviour,Radboud University Nijmegen Medical Centre, The Netherlands

3 Department of Rehabilitation Medicine, Nijmegen Centre for Evidence Based Practice,Radboud University Nijmegen Medical Centre, The Netherlands

4 IQ Scientific Institute for Quality of Healthcare, Nijmegen Centre for Evidence Based Practice,Radboud University Nijmegen Medical Centre, The Netherlands

Correspondence should be addressed to S. H. J. Keus, [email protected]

Received 2 August 2011; Revised 6 October 2011; Accepted 6 October 2011

Academic Editor: Gammon M. Earhart

Copyright © 2012 S. H. J. Keus et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Because of the complex nature of Parkinson’s disease, a wide variety of health professionals are involved in care. Stepwise, wehave addressed the challenges in the provision of multidisciplinary care for this patient group. As a starting point, we havegained detailed insight into the current delivery of allied healthcare, as well as the barriers and facilitators for optimal care.To overcome the identified barriers, a tertiary referral centre was founded; evidence-based guidelines were developed and cost-effectively implemented within regional community networks of specifically trained allied health professionals (the ParkinsonNetconcept). We increasingly use ICT to bind these professional networks together and also to empower and engage patients in makingdecisions about their health. This comprehensive approach is likely to be feasible for other countries as well, so we currentlycollaborate in a European collaboration to improve community care for persons with Parkinson’s disease.

1. Background

The number of patients with Parkinson’s disease (PD) andrelated forms of parkinsonism is increasing in all ageingsocieties [1]. In Western Europe’s five, and the world’s 10most populous nations, the number of individuals with PDover the age of 50 was between 4.1 and 4.6 million in 2005;these numbers will be doubled by 2030 [1]. Likewise, thecosts related to PD care will increase dramatically. PD isa very complex disorder, characterised by a wide array ofboth motor and nonmotor problems for which medical carealone is insufficient [2–6]. As a reflection of this complexity,no less than 18 different disciplines (e.g., physiotherapy andpsychology) may be involved in PD care [7–9]. However,patients often have no access to the allied healthcare required[10]. Moreover, the involvement of various disciplinesrequires close collaboration and integration of medicaland nonmedical care. Great challenges remain in the way

multidisciplinary care is best realized for Parkinson patients.But, where to start? The purpose of this paper is to share thevarious steps we have taken (Figure 1), as they are likely to befeasible for application in other countries.

2. Stepwise Improvement of CommunityHealth Care

Step 1 (gaining insight into current care). Detailed insight intothe current provision of allied healthcare, as well as barriersand facilitators for optimal care, was lacking. As a firststep, we therefore aimed to evaluate current care. Surveysinvolving more than 500 PD patients and 300 allied healthprofessionals were used to gain this insight [11, 12]. Theresults revealed that on average therapists treated as few asthree individual PD patients a year. Therapists also reportedthat they had only limited expertise in treating PD. A majorbarrier for improvement was the absence of guidelines to


8.Continuouseducation

9.

Transparancy7.8.

Supporting communication 6.7.

Training based on guidelines5.6.

Selection of dedicated professionals4.5.

3.Definition of a ParkinsonNet region

4.

Development of evidence-based guidelines2.3.

Creation of a regional expert centre1.2.

1.1.Gaining insight into current care

Figure 1: Steps of the Dutch model to improve community healthcare for Parkinson’s disease.

support these therapists in providing optimal treatment.Most patients were referred by their neurologist. However,referring physicians had no information about the benefitsof, for example, physiotherapy in PD, and were unableto find therapists with PD-specific interest or expertise.Finally, therapists of different disciplines (e.g., speech andlanguage therapy, occupational therapy, and physiotherapy)were often unaware of each other’s treatment possibilities.Moreover, communication about a common patient (e.g.,on treatment goals and timing of interventions) was verypoor. As trends have been found towards positive effectsof integrated care programs in the chronically ill [13] thisneeded to be improved. Therefore, our next step was to createa regional Parkinson, multidisciplinary expert centre.

Step 2 (creation of a regional expert centre). In order tooffer expert care to PD patients and their carers, a regionalParkinson’s expert centre was initiated. The centre serves asa tertiary referral centre for a large catchment area, byoffering critical revision of the diagnosis (if needed sup-ported by ancillary investigations), recommendations with

respect to drug treatment and stereotactic neurosurgery, andindividually tailored multidisciplinary treatment advice. Thecentre also initiates and coordinates clinical trials and dis-seminates the newly acquired knowledge. The centre is partof the neurology department of one of the eight Dutchuniversity medical centres. Several comparable initiativeshave meanwhile arisen across the country.

Step 3 (development of evidence-based guidelines). Next, westarted to develop evidence-based guidelines for alliedhealthcare. As physiotherapy is the most applied alliedhealthcare discipline in PD care, we first developed a guide-line that targeted physiotherapy for PD. Conform interna-tional standards for guideline development, practice recom-mendations were developed based on the results of a sys-tematic literature review, clinical expertise, and patientvalues. The recommendations were graded according tothe level of evidence available [14]. The guideline wasauthorized and distributed by our national professionalorganisation for physiotherapy, the Royal Dutch Society forPhysical Therapy (KNGF). Likewise, guidelines for speech


and language therapy and for occupational therapy weredeveloped, authorized, and distributed [15]. For other disci-plines, for example, dieticians, the development of guidelineshas recently started. The guidelines provide decision supportfor everyday clinical practice.

3. ParkinsonNet

As dissemination does not automatically lead to implementa-tion, we developed a multifaceted implementation strategy:ParkinsonNet. ParkinsonNet not only aims to implementthe guidelines, but also to reorganise allied healthcare toincrease the patient volume of therapists, make experthealthcare professionals visible to other professionals as wellas to patients, and support communication amongst healthcare professionals involved in PD care as well as betweenprofessionals and patients. To set up a ParkinsonNet, first aregion needed to be defined.

Step 4 (definition of a parkinsonNet region). Members of theParkinsonNet team (at the regional expert centre) togetherwith the coordinating neurologist of a general hospital de-fined the catchment area for which the ParkinsonNet neededto be developed. Within this geographic area, a local com-munity of allied health professionals with Parkinson-specificexpertise was created. The key points in this processwere selection, training, communication, and transparency(Figure 1, steps 5 to 8).

Step 5 (selection of dedicated professionals). To succeed in in-creasing the patient volume, a relatively small number oftherapists were selected for each regional ParkinsonNet. Toallow the networks to evolve slowly, the Dutch networkswere set up with a maximum of one physiotherapist forevery 20,000 residents in the specific region, and one speechand occupational therapist for every 40,000 residents. Thesenumbers were based on the preferred patient volume asreported by therapists (i.e., 15), the estimated number of PDpatients in the Netherlands, the current referral pattern ofphysicians (more patients are referred to physiotherapy incomparison with occupational therapy or speech and lan-guage therapy), and the number of residents in the prede-fined region.

For selection, all allied health professionals in a specificregion were informed about the benefits, requirements, andcosts for participation. In all regions, the numbers of phys-iotherapists interested in participation exceeded the requirednumber, making a selection required. Professionals workingin the same neighbourhood were therefore asked to arrangeself-selection. Only when this was not successful, the Par-kinsonNet team made the selection based on motivation,bio sketch and current function in regional PD care. In theselection process we tried to reach a good geographicaldispersion in the region. This, as an evaluation under DutchPD patients and allied health professionals, revealed thatthey both were prepared to travel up to 15 minutes.

Step 6 (training based on guidelines). All allied health pro-fessionals selected for a future network participated in

the same 3-day (for the first networks this was a 4-day),interactive course. Here they were trained to treat PDpatients according to the evidence-based guidelines. Theprogram entailed both mono- and multidisciplinary classes.Participating neurologists and PD nurse specialists wereinformed about the referral criteria and main treatmentoptions of the allied health professions included in their re-gional ParkinsonNet. For physiotherapists, a guideline-basedelectronic patient record was developed to further supporttheir clinical decision taking. During the course, physiother-apists were trained to use this patient record.

Step 7 (supporting communication). Starting at the course,networking was supported. For example, professionals froma specific region shared a table during lunch and togetherprepared and completed educational tasks during the course.After the course, continuation of network meetings takesplace during three-monthly regional seminars and a yearlyParkinsonNet congress (see Step 9). Concerning communi-cation about common patients, the development of regionalcommunication plan was facilitated. In addition, a securedweb-based community is used to enhance communication,both within the ParkinsonNet as with hospital professionals.

Step 8 (transparency). Patients, medical and nonmedical careprofessionals were informed about the ParkinsonNet. Thelocation of the specialized therapists is visualized by web-based sources and printed folders. Moreover, structured andpreferred referral to ParkinsonNet therapists by neurologistswas supported by using standardized referral forms, includ-ing objective referral criteria. So, participating therapist wereenabled to attract a large number of patients. A certificationsystem, supported by professional societies, was developedto guarantee the quality of therapy provided by healthprofessionals participating in these networks.

Step 9 (continuous education and exchange of knowledge).Each regional ParkinsonNet organises, if needed with sup-port by the ParkinsonNet team, three-monthly seminars tofor example, practice skills, discuss cases, and to enhance(multidisciplinary) collaboration within the region. Everyyear, the ParkinsonNet team organises a national Parkinson-Net congress with national and international speakers and awide variety of Parkinson’s related workshops (e.g., on cogni-tive functioning, sleep, nutrition, and exercise and the brain).In addition, through a secured web-based community, up-to-date information is shared by the ParkinsonNet team.

4. Scientific Evaluation ParkinsonNet

In 2004, the first, multidisciplinary ParkinsonNet was de-signed and tested for its feasibility [17]. This ParkinsonNetincluded neurologists, a PD nurse specialist, physiothera-pists, occupational therapists, and speech and language ther-apists. Given its feasibility and the enthusiasm of the par-ticipants, a cluster-randomized trial was designed to furtherevaluate ParkinsonNet. For feasibility purposes of the trial,eight ParkinsonNets were developed which only includedthe most used allied healthcare in PD, that is, physiotherapy.


2

3

4

51

116 7

810

29

24

27

2625

232221

18 191617

15

13 14

12

65

2830

31

32 33

34

35

47454644

42 4341

40 64

37

36

38

20

39

48

49 50

51

56

57585960

63

62

52

53

5455

9

Figure 2: Coverage of ParkinsonNet throughout the Netherlands.

Theses clusters were compared with eight clusters where careremained unchanged. The trial, in which 699 PD patientsparticipated, showed that the quality of care increased andthe volume of patients per therapist more than doubledwithin as little as six months while considerably savingcosts [18]. In addition, evaluation of the connectedness ofhealthcare professionals within the ParkinsonNet showedthat especially therapists treating more than nine PD patientsa year were associated with stronger connectedness withother health professionals than those treating less than 10PD patients a year. As connectedness between professionalsis known to influence clinical decision making and thecoordination of patient care [19], this knowledge is of highimportance to the size of future networks.

5. National Coverage ParkinsonNet

Supported by these positive results, ParkinsonNet was en-dorsed by professional healthcare organizations and thenational patient society. In 2010 national coverage within theNetherlands was achieved by 65 unique networks (Figure 2).The size of the networks is related to the population densityof the specific regions. In addition to increase in number ofnetworks, many additional disciplines have been added to theParkinsonNet. Currently, throughout the Netherlands, 1885

care professionals are participating, amongst which 57 neu-rologists, 107 PD nurse specialists, 809 physiotherapists, 317occupational therapists, 318 speech and language therapists,89 dieticians, 76 elderly care physicians, 62 psychologists, 31social workers, and 4 sex therapists.

6. Multidisciplinary Guidelines

In addition to the monodisciplinary guidelines, a multidisci-plinary guideline has been developed in a joint collaborationamong professional organizations of 18 medical professions,the patient society, and two national healthcare knowledgeand quality institutes [7]. The multidisciplinary guidelineincludes recommendations not only for daily medical andnonmedical practice, as an update of the NICE guidelines[8], but also for network care. Specifically these recommen-dations, concerning collaboration, expertise, communica-tion, and finances were lacking in the existing guidelines [20],even though they are of high importance. For example, inoutpatient neurology, dissatisfaction with communication isrelated to noncompliance [21]. As part of the support forcollaboration, the guideline provides a detailed overview ofimpairments, limitations, restrictions, and external factorsrelated to PD (Figure 3) [7]. For this overview the commonlanguage of the International Classification of Functioning,


Dysfunction of the basal ganglia, caused by degeneration of dopamine-producing cells in the substantia nigra (ICD-10: G20)

Impairments in functions Activity limitations and restrictions inparticipation:

b1: Mental functionsDementia (b117)Impairments in temperament and personality (b126)Impairments in energy and drive functions, e.g. reduced motivation

Sleep impairments (b134)Reduced attention (b140)

Impairments in perceptual functions, e.g. reduced visuospatial perception and hallucinations∗

Impairments in higher level cognitive functions, e.g. in planning, decision-making and mental flexibility (b164)Impairments in mental functions of language, e.g. verbal perseveration (b167)

b2: Sensory functions and painSeeing impairments, e.g. visual acuity∗ (b210)Dizziness∗ (b240)∗

Impairments in smell (b255)Proprioceptive function (b260)Tingling (b265)(Central) pain (b280)

b3: Voice and speech functions (b3)Reduced pitch and loudness of voice (b310)Impaired articulation (including dysarthria) (b320)Reduced fluency of speech (b330)

b4: Functions of the cardiovascular and respiratory systemsImpairments in blood pressure (e.g. orthostatic hypotension∗)(b420)

Reduced exercise tolerance∗ (b455)

b5: Functions of the digestive system

swallowing (b510) Constipation∗ (b525)Reduced weight maintenance (b530)

b6: Genitourinary and reproductive functionsImpaired urination, e.g. (urge)incontinence∗ (b620)Impaired sexual functions, e.g. impotence and increased sexual

b7: Neuromusculoskeletal and movement-related functionsReduced joint mobility∗ (b710)Reduced muscle power∗ (b730)Impaired muscle tone functions, e.g. rigidity and dystonia (b735)Reduced muscle endurance∗ (b740)Impaired motor reflex functions (b750), e.g. simultaneous contraction of antagonistsReduced postural responses (b755)Reduced control of voluntary movements (b760), e.g. dysdiadochokinesia, reduced ‘’motor set’’ causing starting problems and reduced or absence of internal cues causing problems in automated, sequential movementsImpaired involuntary movement functions (b765), e.g. bradykinesia, (resting) temor and dyskinesia∗

Impairments in gait patterns, e.g. asymmetry, freezing, reduced step length, velocity, trunk rotation and arm swing (b770)

d1: Learning and applying knowledge, e.g.acquiring skills (d155), writing (d170),solving problems (d175) and makingdecisions (d177)

d2: General tasks and demands, e.g..undertaking multiple tasks (d220), carryingout daily routine (230), handling stress andother psychological demands (d240) d3: Communication, e.g. speaking (d330), producing non-verbal messages (d335),writing messages (d345)

d4: Mobility, e.g. changing and maintainingbody position (d410–d429), carrying, moving and handling objects (d430–d449), walkingand moving (d450–d469), moving aroundusing transportation (d470–d489) d5: Self-care, e.g. washing oneself (d510),toileting (d530), dressing (d540), eating(d550) and drinking (d560) d6: Domestic life, e.g. preparing meals (d630) and doing housework (d640)d7: Interpersonal interactions andrelationships , e.g. basic interpersonalinteractions (d710) and particularinterpersonal relationships with strangers,formal persons, family and husband or wife (d730–d779)

d8: Major life areas, e.g. education (d810–839), work and employment (d840–d859)and economic life (d860–d879)

d9: Community, social and civic life , e.g.community life (d910), recreation andleisure (d920), religion (d930) and politicallife (d950)

Environmental factors, e.g.: Personal factors, e.g.:

e1: Products en technology, e.g. assistive devices, design of homes and public buildings and financial assetse2: Natural environment and human-made changes to environment, e.g. population density, flora, fauna, climate, light intensity e3: Support and relationships, e.g. with family, friends, colleagues, people in the authority and health professionalse4: Attitudes, e.g. of people, social attitudes and norms e5: Services, systems and policies, e.g. housing, transportation, communication, social support, health services and education

Age and genderEducationExperiences and preferencesCo-morbidity and coping skills

These are not ICF classified because of large social and cultural variances

On/off periods∗

interest∗ (b640)

and impulse control∗ (b130)

Impairments in emotion, e.g. anxiety∗ (b152)

•••••••

•

•

•••••

•

•

•

•

•

•

•

•

•

•••••••

••

••

•

•

•

•

••

•

•••

••

••

••

•

••••

b8: Functions of the skin and related structuresImpairments in sweating and sebum production (b830)Impaired sensations related to the skin (pins and needles) (b840)

(b156)

Impaired ingestion, e.g. drooling, vomiting∗ and impaired

(b798)

Figure 3: ICF model for Parkinson’s disease. ICF: International Classification of Functioning, Disability and Health; numbers, for example,b770, refer to codes in the ICF Handbook [16].

Disability, and Health (ICF) was used [16]. So far, an ICFfor neurology combining three neurologic disorders wasavailable, but not for PD specific [22]. This PD-specific clas-sification can further improve communication in relation topatient functioning between health care workers, researchers,and social policy makers.

7. Empowering Patients

Traditionally, the relationship between patients and theirhealthcare providers is fairly paternalistic, with healthcareproviders making decisions (with the best intentions) andpatients simply carrying out instructions. Some patients ap-preciate this role. However, many patients wish to have morecontrol over their own care [23] and patient preferencesmay differ from what doctors focus on [24]. The guidelinesupports patient empowerment, for example, by teachingtherapists how to get to the treatment goal in “partnership”with the patient. This, however, will not be sufficient [25].Increasingly, the Internet can provide solutions to supportpatient empowerment [26]. To further support patientswithin ParkinsonNet regions to participate in medicaldecisions made about their health, we developed a web-based “portal to empower patients.” The portal providespatients with information necessary to make choices in theirown health care process. For example, this includes actual

and controlled information about all treatment options. Inaddition, patients are enabled to easily find a specializedhealth care professional within their community, based ontransparent background information (e.g., the number of PDpatients treated by a professional, or the education receivedto increase Parkinson-specific knowledge and expertise).Another tool for patients will be the opportunity to buildtheir own virtual network of care providers, supportinginformation sharing and collaboration between all partici-pants.

8. International Collaboration

As described, one of our first steps to improve PD carewas the development of an evidence-based physiotherapyguideline [14]. An external quality evaluation of all Parkin-son guidelines available worldwide, by the Dutch Institutefor Health Care Improvement (CBO), showed that thisguideline is one of the few which is of good quality [7].In addition, to date, it is still unique in its field. As aconsequence, internationally there is a lot of interest forusing the guideline. The Association for Physiotherapists inParkinson’s Disease Europe (APPDE; http://www.appde.eu/)therefore endorses the guideline and its implementation.

Currently, we are updating the guideline, in a jointcollaboration among 19 European physiotherapy associa-tions, members of the European Region of the World


Confederation for Physical Therapy (ER-WCPT). This willlead to a first European guideline for physiotherapy in Par-kinson’s disease in 2012. As a first step in the guideline de-velopment, surveys have been set out Europe-wide (n =10, 000) to gain insight into current care, barriers met indelivering this care, possibilities for improvements and toidentify those therapists interested to become members ofa PD expert’s network. Perhaps this is another first steptowards improving community healthcare (Figure 1). Theresults of the survey will be used to develop key questionsfor the European guideline. In addition, the results will sup-port the participating countries to further structure theirguideline implementation plans.

At the same time, possibilities for using the ParkinsonNetconcept for implementation of the guideline are being ex-plored in several European countries as well as in the Unit-ed States. The approach seems applicable to other countrieswith a similar population density and health care system(e.g., compensation of physical therapy). Through these fu-ture networks, also the guidelines for occupational ther-apy and for into English in collaboration with the NationalParkinson Foundation (http://www.parkinson.org/), can beimplemented and thus leading to multidisciplinary Parkin-son’s networks in many countries.

9. Conclusions

Given the complex nature of PD, many disciplines will beinvolved in Parkinson care. Aiming for optimal Parkinsoncare, in the Netherlands, care for persons with Parkinson’shas been changed stepwise. The Dutch approach seems ap-plicable to other countries, be it with adaptations based onpopulation density and health care organization.

References

[1] E. R. Dorsey, R. Constantinescu, J. P. Thompson et al., “Pro-jected number of people with Parkinson disease in the mostpopulous nations, 2005 through 2030,” Neurology, vol. 68, no.5, pp. 384–386, 2007.

[2] G. Alves, T. Wentzel-Larsen, D. Aarsland, and J. P. Larsen,“Progression of motor impairment and disability in Parkinsondisease: a population-based study,” Neurology, vol. 65, no. 9,pp. 1436–1441, 2005.

[3] J. Jankovic, M. McDermott, J. Carter et al., “Variable expres-sion of Parkinson’s disease: a base-line analysis of the DATA-TOP cohort,” Neurology, vol. 40, no. 10, pp. 1529–1534, 1990.

[4] D. Muslimovic, B. Schmand, J. D. Speelman, and R. J. DeHaan, “Course of cognitive decline in Parkinson’s disease: ameta-analysis,” Journal of the International NeuropsychologicalSociety, vol. 13, no. 6, pp. 920–932, 2007.

[5] B. Post, J. D. Speelman, and R. J. de Haan, “Clinical heter-ogeneity in newly diagnosed Parkinson’s disease,” Journal ofNeurology, vol. 255, no. 5, pp. 716–722, 2008.

[6] K. R. Chaudhuri and P. Odin, “The challenge of non-motorsymptoms in Parkinson’s disease,” Progress in Brain Research,vol. 184, no. C, pp. 325–341, 2010.

[7] B. R. Bloem, T. van Laar, S. H. J. Keus et al., MultidisciplinairyGuideline “Parkinson’s Disease”, Van Zuiden Communications,Alphen aan den Rijn, The Netherlands, 2010.

[8] NICE, Parkinson’s Disease. Diagnosis and Management in Pri-mary and Secondary Care, NICE Clinical Guideline 35, Na-tional Collaborating Centre for Chronic Conditions, London,UK, 2006.

[9] M. A. van der Marck, J. G. Kalf, I. H. W. M. Sturkenboom, M.J. Nijkrake, M. Munneke, and B. R. Bloem, “Multidisciplinarycare for patients with Parkinson’s disease,” Parkinsonism andRelated Disorders, vol. 15, no. 3, pp. S219–S223, 2009.

[10] Parkinson’s Audit 2009 Report, Parkinson’s UK, London, UK.[11] S. H. J. Keus, B. R. Bloem, D. Verbaan et al., “Physiotherapy in

Parkinson’s disease: utilisation and patient satisfaction,” Jour-nal of Neurology, vol. 251, no. 6, pp. 680–687, 2004.

[12] M. J. Nijkrake, S. H. J. Keus, R. A. B. Oostendorp et al., “Alliedhealth care in Parkinson’s disease: referral, consultation, andprofessional expertise,” Movement Disorders, vol. 24, no. 2, pp.282–286, 2009.

[13] M. Ouwens, H. Wollersheim, R. Hermens, M. Hulscher, and R.Grol, “Integrated care programmes for chronically ill patients:a review of systemic reviews,” International Journal for Qualityin Health Care, vol. 17, no. 2, pp. 141–146, 2005.

[14] S. H. J. Keus, B. R. Bloem, E. J. M. Hendriks, A. B. Bredero-Cohen, and M. Munneke, “Evidence-based analysis of physicaltherapy in Parkinson’s disease with recommendations forpractice and research,” Movement Disorders, vol. 22, no. 4, pp.451–460, 2007.

[15] J. G. Kalf, I. Sturkenboom, M. Thijssen, B. J. M. de Swart, B.R. Bloem, and M. Munneke, “New guidelines in Parkinson’sdisease: occupational therapy and speech therapy,” EPNNJournal, vol. 4, pp. 12–15, 2008.

[16] World Health Organization (WHO), “International Classifica-tion of Functioning, Disability and Health (ICF),” http://apps.who.int/classifications/icfbrowser/.

[17] M. J. Nijkrake, S. H. Keus, S. Overeem et al., “The Parkinson-Net concept: development, implementation and initial experi-ence,” Movement Disorders, vol. 25, no. 7, pp. 823–829, 2010.

[18] M. Munneke, M. J. Nijkrake, S. H. Keus et al., “Efficacy of com-munity-based physiotherapy networks for patients with Par-kinson’s disease: a cluster-randomised trial,” The LancetNeurology, vol. 9, no. 1, pp. 46–54, 2010.

[19] M. Wensing, M. Van der Eijk, J. Koetsenruijter, B. R. Bloem, M.Munneke, and M. Faber, “Connectedness of healthcare profes-sionals involved in the treatment of patients with Parkinson’sdisease: a social networks study,” Implementation Science, vol.6, no. 1, p. 67, 2011.

[20] M. Fargel, B. Grobe, E. Oesterle, C. Hastedt, and M. Rupp,“Treatment of Parkinson’s disease: A survey of patients andneurologists,” Clinical Drug Investigation, vol. 27, no. 3, pp.207–218, 2007.

[21] K. A. Grosset and D. G. Grosset, “Patient-perceived involve-ment and satisfaction in Parkinson’s disease: Effect on therapydecisions and quality of life,” Movement Disorders, vol. 20, no.5, pp. 616–619, 2005.

[22] M. Leonardi, P. Meucci, D. Ajovalasit et al., “ICF in neu-rology: functioning and disability in patients with migraine,myasthenia gravis and Parkinson’s disease,” Disability andRehabilitation, vol. 31, no. 1, pp. S88–S99, 2009.

[23] M. van der Eijk, M. J. Faber, S. Al Shamma, M. Munneke, andB. R. Bloem, “Moving towards patient-centered healthcare forpatients with Parkinson’s disease,” Parkinsonism and RelatedDisorders, vol. 17, no. 5, pp. 360–364, 2011.

[24] L. J. Findley and M. G. Baker, “Treating neurodegenerative dis-eases,” BMJ, vol. 324, no. 7352, pp. 1466–1467, 2002.


[25] D. M. Berwick, “What “patient-centered” should mean: con-fessions of an extremist,” Health Affairs, vol. 28, no. 4, pp.w555–w565, 2009.

[26] L. Fincher, C. Ward, V. Dawkins, V. Magee, and P. Willson,“Using telehealth to educate Parkinson’s disease patients aboutcomplicated medication regimens,” Journal of GerontologicalNursing, vol. 35, no. 2, pp. 16–24, 2009.


Research Article

Altered Dynamic Postural Control during Step Turning inPersons with Early-Stage Parkinson’s Disease

Jooeun Song,1 Susan Sigward,1 Beth Fisher,1, 2 and George J. Salem1, 2

1 Jacquelin Perry Musculoskeletal Biomechanics Research Laboratory, Division of Biokinesology and Physical Therapy,University of Southern California, Los Angeles, CA 90089-9006, USA

2 Department of Neurology, Keck School of Medicine, University of Southern California, Los Angeles, CA 90089-9006, USA

Correspondence should be addressed to Beth Fisher, [email protected]

Received 31 July 2011; Revised 26 September 2011; Accepted 17 October 2011


Copyright © 2012 Jooeun Song et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Persons with early-stage Parkinson’s disease (EPD) do not typically experience marked functional deficits but may have difficultywith turning tasks. Studies evaluating turning have focused on individuals in advanced stages of the disease. The purpose of thisstudy was to compare postural control strategies adopted during turning in persons with EPD to those used by healthy control(HC) subjects. Fifteen persons with EPD, diagnosed within 3 years, and 10 HC participated. Participants walked 4 meters and thenturned 90◦. Dynamic postural control was quantified as the distance between the center of pressure (COP) and the extrapolatedcenter of mass (eCOM). Individuals with EPD demonstrated significantly shorter COP-eCOM distances compared to HC. Thesefindings suggest that dynamic postural control during turning is altered even in the early stages of PD.

1. Introduction

Postural control is the ability to alter the magnitude andpatterns of segmental kinematics (e.g., trunk and limbmovements) in order to direct body position in responseto external mechanical demands imposed during static anddynamic tasks such as turning [1, 2]. Functional indepen-dence, and consequently quality of life, is compromised inindividuals with postural control deficits. Persons with early-stage Parkinson’s disease (EPD), Hoehn and Yahr stage 1and 2, may not demonstrate overt clinical symptoms andmay describe only minimal levels of functional impairment,such as reduced gait velocity and stride length, duringsimple movement tasks including straight walking [3, 4].However, they often demonstrate altered postural controlduring standing tasks [5] and report difficulty with turning[6]. Turning difficulty becomes a sensitive indicator of ahigher prevalence of freezing and falling in persons withadvanced PD (Hoehn and Yahr stage ≥3 with moderate tosevere symptoms) [7, 8].

Turning is a challenging task that aims to transport thebody’s mass in a new direction. It requires deceleration of thebody’s center of mass (COM; the position that represents the

equilibrium point of the body’s mass), rotation of the axialsegments, and acceleration of the COM in the new direction[9, 10]. This is accomplished in three consecutive steps:approach, pivot, and acceleration steps embedded withintwo phases. During these phases individuals transition fromdouble limb to single limb stance before returning to doublelimb stance [11]. Studies have established that in young,healthy individuals the redirection of the COM into thenew direction of travel is initiated through appropriate footplacement during phase 1 (from approach step to pivot step).During phase 2 (from pivot step to acceleration step) trunkmovements are used to control turning [12].

The demands of turning present unique challenges toindividuals with impaired postural control as they arerequired to initiate a state of disequilibrium during singlelimb stance in order to change directions during an ongoingmovement [9, 13]. This disequilibrium is created by increas-ing the distance between the body’s COM and the center ofpressure (COP; the equilibrium point of the distribution ofthe resultant ground reaction force applied to the base ofsupport). An increased distance between these two pointsnot only creates momentum necessary to turn, but alsorequires increased neuromuscular control (e.g., neural drive,


Table 1: Mean and standard deviation of participant’s characteristics.

Group

EPD (n = 15) HC (n = 10) Differences (95% CI)

Age (yr) 62 (9.1) 60 (8.5) 2 (−5.49; 9.49)

Height (m) 1.68 (0.07) 1.72 (0.09) −0.04 (−0.11; 0.03)

Weight (kg) 68.9 (12.1) 74.8 (17.2) −5.9 (−17.99; 6.19)

Approach gait velocity (m/s) 1.35 (0.14) 1.46 (0.14) −0.11 (−0.23; 0.01)

Mean (Standard Deviation).EPD: persons with early-stage Parkinson’s disease.HC: healthy control participants.

muscle forces, and joint power) to redirect and control thismomentum. Alterations in turning strategies are thoughtto reflect an individual’s inability to meet these increasedneuromuscular demands. For example, when comparedto healthy controls, persons diagnosed with advanced PDutilize postural control strategies that include longer turningtimes [14] along with a greater number of smaller steps[8, 14] to complete a turn. These postural adjustmentsserve to decrease the body’s momentum, reduce the distancebetween the COM and the COP, and in turn decreasethe neuromuscular demands. While alterations in posturalcontrol strategies have been observed in individuals withadvanced PD, they have not been characterized in individualsdiagnosed with EPD.

Individuals diagnosed with EPD do report difficultyturning [6]. However, in contrast to individuals diagnosedwith advanced PD, they do not frequently exhibit observablemovement impairments that could impact turning such asshuffling gait, freezing episodes, and en bloc movements.[3, 4, 15, 16]. A more detailed evaluation of posturalcontrol strategies employed by individuals diagnosed withEPD during turning is needed. However, more traditionalmeasures of postural control that relate the distance betweenthe positions of the COP and COM may not be sensitiveenough to detect differences between individuals diagnosedwith EPD and healthy controls because they do not takeinto account the dynamic nature of the turning task. Duringdynamic tasks it is important to consider not only theposition of the COM but also the magnitude and directionof the COM velocity in relation to the COP [17, 18].

Despite self-reports of difficulty turning in persons withEPD, studies to date have not characterized the posturalcontrol strategies used during turning in this cohort. Earlyidentification of these strategies may be used to developeffective intervention protocols that (1) improve turningcapabilities and (2) increase balance confidence in personswith EPD. Therefore, the purpose of this study was tocharacterize the differences in postural control during a stepturning activity, between persons with EPD and healthyage-matched control (HC) participants. We hypothesizedthat, compared to HC participants, persons with EPDwould demonstrate a dynamic postural control strategythat reduced the demands on the neuromuscular system.Specifically, we hypothesized that when accounting for theposition, magnitude and velocity of the COM persons withEPD would demonstrate shorter distances between the COP

and the eCOM than healthy controls during both phases of astep turn at 90 degrees. Moreover, this will be accomplishedby both decreasing their COM velocity and the distancebetween their COP and the COM.

2. Methods

2.1. Participants. Fifteen persons with EPD and 10 HC sub-jects participated. Participant characteristics are provided inTable 1. A fellowship-trained movement disorder specialistconfirmed diagnosis of idiopathic PD in our participants,performed the Unified Parkinson’s Disease Rating Scale(UPDRS), and determined Hoehn and Yahr stage for eachindividual participant. Participants that had pharmacologi-cal treatment were stable and tested while they were on theirroutine therapy (Table 2). At the time of testing, none ofthe participants exhibited any fluctuations in motor abilitythroughout the day, dyskinesia, dystonia, or other signs ofinvoluntary movement.

The inclusion criteria for the early PD group were thefollowing: (1) age ≥18 years old, (2) able to ambulate atleast 14 meters (time not measured) without a walker orother devices, (3) diagnosed with PD within 3 years [19],(4) Hoehn and Yahr stages 1-2 (indicating EPD), and (5)stable on PD medications. Healthy control participants wereage and gender-matched to the participants in the early PDcohort. Participants were excluded from the study for thefollowing: (1) surgical intervention for persons with PD, (2)Mini-Mental State Exam (MMSE) score <24 [20], (3) comor-bidities affecting gait (e.g., diabetes, musculoskeletal injury,arthritis, vestibular disorders), (4) severe vision problems,and (5) pregnancy.

2.2. Protocol. All testing took place in the MusculoskeletalBiomechanics Research Laboratory at the University ofSouthern California (USC). Procedures were explained toeach participant and each participant signed an informedconsent form approved by the Institutional Review Board ofthe USC. Participants were instructed to walk straight at a“self-selected, comfortable pace and turn at the designatedstanchions at a right angle” toward their dominant leg andthen continue walking in the new direction (Figure 1).Prior to testing the dominant leg was determined as the legthey would use to kick a ball as far as possible. No otherinstructions were given to the participants. The subjects


Table 2: Dosage of Parkinson’s medications.

Participant number Medication Dosage Frequency

P01 Levodopa/carbidopa 25–100 mg 3x/day

P02 De Novo

P03Pramipexole 1 mg 3x/day

Rasagiline 1 mg 1x/day

Levodopa/carbidopa 50–200 mg 3x/day

P04 De Novo

P05Pramipexole 1.5 mg 3x/day

Selegiline 5 mg 2x/day

P06Rasagiline 1 mg 1x/day

Amantadine 100 mg 2x/day

P07 De Novo

P08Rasagiline 1 mg 1x/day

Trihexyphenidyl 4 to 6 mg 1x/day

P09 Levodopa/carbidopa 150 mg 3x/day





P12 Levodopa/carbidopa 25–100 mg 4x/day

P13Levodopa/carbidopa 25–100 mg 3x/day

Rasagiline 0.5 mg 1x/day

P14 Pramipexole 0.5 mg 3x/day


Selegiline 5 mg 2x/day

Trihexyphenidyl 2 mg 3x/day

A

B

Starting point

2.4 m

1 m

4 m

0.6 m

Figure 1: Laboratory setup. Dashed line: starting point; A: first trigger; B: second trigger; black square: force plate (AMTI 1.2 m × 1.2 m,1560 Hz). Two stanchions were placed at the midpoint of each force plate. Starting point to A: 1 m. A to B: 2.4 m. B to force plate: 0.6 m.Force plate to stanchions: 0.6 m.


started 1 meter from the first timing trigger. They thenwalked for an additional 2.4 meters before walking throughthe second timing trigger, which was located 0.6 meters infront of the force plate.

A total of 10 turning trials were recorded for eachparticipant. The first three successful trials during which theyused a step turn strategy were considered for analysis. A stepturn is defined as a change in direction opposite to the pivotfoot [10, 13]. Ninety-degree turns were selected for analysisbecause these types of turns are associated with the naviga-tion of corridors, street corners, and other common walkingactivities. Moreover, Sedgman and colleagues reported thatthe majority of turns experienced during activities of dailyliving were between 76◦ and 120◦ [21].

Kinematic data were sampled at 60 Hz using a motionanalysis system (Vicon 612, Oxford Metrics Ltd., Oxford,England). Reflective markers (14 mm spheres) were placedbilaterally on the skin over specific anatomical landmarksincluding the anterior, posterior, and lateral cranium, acro-mion processes, anterior and posterior shoulders, greatertubercles of humerus, medial and lateral humeral epi-condyles, radial styloid process, ulnar head, third metacar-pophalangeal joints, 7th cervical vertebrae, sternoclavicularnotch, iliac crests, anterior superior iliac spines, poste-rior superior iliac spines, L5-S1 joint, medial and lateralfemoral epicondyles, medial and lateral malleoli, first andfifth metatarsal heads, and first proximal/distal phalanx.Additionally, cluster markers were placed with a band overthe upper arms, lower arms, thighs, shanks, and shoeheels. Reflective markers were identified manually withinthe VICON Workstation software and then imported intoVisual 3D software (C-Motion, Rockville, MD). 3D markercoordinates were lowpass filtered at a cut-off frequency of6 Hz.

Kinetic data were captured using 1.2 m × 1.2 m AMTI(Advanced Mechanical Technologies, Inc., Newton, MA,USA) force platform at 1560 Hz. The size of the plat-form allowed for quantification of ground reaction forcesthroughout the entire task. Kinematic and kinetic datawere interfaced to the same microcomputer allowing forsynchronization of data.

2.3. Data Analysis. Dynamic postural control during turningwas quantified using the method previously described byHof (1) [17]. It was defined as the difference between theCOP and an extrapolated COM calculated to account for theposition, magnitude and velocity of the COM:

Dynamic Postural Control = COP−(

COG +COMvel√(

g/l)).

(1)

The COP was determined from the forces and momentsobtained from the force platform. The position of thetotal body COM was defined using the weighted sum ofthe COM of all 15-body segments. Based on Winter [22],instantaneous velocity of the total body COM (COMvel)

Approachstep

Pivotstep

Accelerationstep

Phase 2

Ph

ase

1

StanchionsStanchions

Figure 2: Schematic representation of the COP and the eCOMtrajectories during the step turn. Solid line: eCOM trajectory;dashed gray line: COP trajectory; Phase 1: from approach step topivot step; Phase 2: from pivot step to acceleration step.

was computed from the linear total body COM positions(COMpos):

COMveln =[COMposn + 1− COMposn− 1

]Δt

, (2)

where, n is the event frame, and Δt is the time between eventframes.

The center of gravity (COG) represents the verticalprojection of the body’s COM. It was calculated based onthe medial-lateral and the anterior-posterior locations ofthe COM. The COM velocity was divided by the naturalfrequency of the limb. The natural frequency was calculatedas√

(g/l) where g is the acceleration of gravity and l is thelength of the leg measured from the ankle joint center tothe COM. The extrapolated COM (eCOM) was calculatedas sum of the COG and the new COM velocity term:

eCOM =(

COG +COMvel√(

g/l)). (3)

The turning cycle was defined from heel strike of theapproach step to heel strike of the acceleration step and wasbroken into 2 phases. Phase 1 was defined from heel strike ofthe approach step to heel strike of the pivot step. Phase 2 wasdefined from heel strike of the pivot step to heel strike of theacceleration step (Figure 2). The dependent variable dynamicpostural control was measured as the peak distance betweenthe COP and the eCOM. Peak distance between the COPand the COG, and the peak COM velocity were identifiedfor each phase. These measures were considered in the casein which dynamic postural control differed between groups,as alterations in both position and velocity can affect thismeasure of dynamic postural control. The average approachgait velocity across three successful trials was calculatedover the 2.4 meters between the first trigger (A) and thesecond trigger (B) during turning (Figure 1). The single- anddouble-limb gait cycle phases were determined using forceplate contact and the vertical velocity of the virtual center ofeach foot [23].

2.4. Statistical Analysis. To determine if differences inour dependent variable, dynamic postural control, existed


0

0.1

0.2

0.3

0.4

0.5

0.6

1 11 21 31 41 51 61 71 81 91

Dis

tan

ce(m

)

HS TO HS TOPhase1 Phase2 HS

100% turning cycle

∗

††

(a) Resultant COP-eCOM distance

0

0.1

0.2

0.3

0.4

0.5

0.6

1 11 21 31 41 51 61 71 81 91

Dis

tan

ce(m

)


100% turning cycle−1

†

†

(b) Resultant COP-COG distance

1 11 21 31 41 51 61 71 81 91


100% turning cycle

0

0.2

0.4

0.6

0.8

1

1.2

Vel

ocit

y(m

/s)

††

(c) Resultant COM velocity

Figure 3: Resultant COP-eCOM distance (a), and resultant COP-COG distance (b), resultant COM velocity (c) between groups during thestep turn cycle (±sd). HS: heel strike; TO: toe off; light gray line: persons with early-stage Parkinson’s disease; black line: healthy controlparticipants; gray shadow: double limb support time; white shadow: single limb support time. †denotes statistically significant differencebetween groups (P < 0.05). ∗denotes statistically significant difference between phases (P < 0.05).

between persons with EPD and HC participants acrossturning phases, a 2 × 2 (group × phase) ANOVA wasperformed. In the case in which differences in dynamicpostural control were found between groups, independentt-tests were performed to determine if group differencesexisted in the input variables used to calculate dynamicpostural control, position of the COG relative to the COP,and the COM velocity within each phase. All statisticalanalyses were performed using SPSS 15.0 (Chicago, IL) withan alpha level set at 0.05.

3. Results

Participant characteristics and approach gait velocity areprovided in Table 1. There were no significant groupdifferences for age, height, weight, or approach gait velocity(P > 0.05). In the EPD group, average time since diagnosiswas 18.2 ± 13.9 months, average H&Y score was 1.9 ± 0.3,and average UPDRS motor score was 21.2 ± 6.7. Average

UPDRS gait and postural stability subscores were 0.1 ± 0.4and 0.3± 0.5, respectively.

No significant group by phase interaction was found fordynamic postural control (F = 0.584, P = 0.453). Maineffects of group and phase are found for dynamic posturalcontrol. Persons with EPD demonstrated statistically signif-icant smaller peak COP-eCOM distances compared to HCparticipants during both Phase 1 (20.6% difference; 0.34 ±0.05 versus 0.41 ± 0.06 m; P < 0.01) and Phase 2 (21.1%difference; 0.38± 0.06 versus 0.46± 0.07 m; P = 0.01) of thestep turn (Figure 3(a)). The peak distance between the COPand the eCOM always occurred during single limb stancewithin each of the phases.

Compared to control participants, persons with EPDdemonstrated statistically significant smaller peak COP-COG distances during both Phase 1 (30.8% difference; 0.13±0.03 versus 0.17 ± 0.03 m; P < 0.01) and Phase 2 (28.6%difference; 0.21 ± 0.05 versus 0.27 ± 0.04 m; P < 0.05;Figure 3(b)).


Although there was no significant difference in theaverage approach gait velocity between groups, persons withEPD exhibited significantly slower peak COM velocity whencompared with control participants during phase 1 (14%difference; 0.64 ± 0.10 versus 0.74 ± 0.10 m/s) and phase 2(35% difference; 0.41± 0.11 versus 0.63± 0.08 m/s; P < 0.05;Figure 3(c)) of the turning cycle.

4. Discussion

This study identified differences in dynamic postural controlstrategies in persons with EPD during step turning activitiescompared to HC participants. Using the eCOM, we were ableto account for not only the position of the COM but alsothe magnitude and velocity. This is particularly importantduring turning as the momentum of the COM is needed forforward propulsion and redirection. We found that personswith EPD utilized shorter distances between the COP andthe eCOM during both phases of the turning cycle. Forboth phases the group differences were noted during singlelimb stance. This suggests that during a time in whichpostural control demands are greatest, individuals with EPDadopt a strategy that aims to decrease these demands. Itis also important to note that the differences observed inpostural control between the groups appear to be largelydriven by alterations in magnitude, not timing, suggestingthat individuals with EPD are not adopting an entirely newstrategy but merely scaling the strategy typically used to turn(Figure 3).

Persons with EPD appeared to scale both position andvelocity of the COM: factors used to calculate posturalcontrol in this study. Both the shorter peak distance betweenthe COP and the COG, and the slower peak COM velocityexhibited by individuals with EPD during the turn limit thedisequilibrium experienced by the individual. Both of theseadjustments have the potential to decrease neuromusculardemands. A smaller COP-COG distance reduces the momentarm created for the body weight vector acting aroundthe centers of joint rotation, and thus the magnitude ofthe muscular force required to control the COM [24].Additionally, a slower COM velocity reduces the momentumof the COM and decreases the muscular force required todecelerate and redirect the COM.

These alterations are consistent with what has beenobserved in persons with more advanced PD (longer turningtime and smaller steps) during turning [8, 14, 25]. Thecurrent data support self-reports of “difficulty in turning”from persons with EPD [6]. Our findings are similar tothose reported by researchers investigating other transitionalmovement patterns in persons with more advanced PD,namely gait initiation and sit-to-walk activities. For example,Martin and colleagues [26] reported a shortening of theseparation between the COM and the COP during gaitinitiation in persons with PD, compared to healthy olderadults. Moreover, Buckley and colleagues [27] reportedthat compared to healthy control subjects, persons withPD utilized a conservative movement strategy that limitedseparation of COP-COM during sit-to-walk transitions. Wedemonstrated that when we challenged individuals with EPD

with a turning task, alterations in postural control similar tothose seen in more advanced stages were observed. This is ofparticular importance since this group does not commonlydemonstrate obvious signs of gait disturbance [3, 4].

Although the current findings describe adjustments inpostural control in individuals with EPD, they are notsufficient to tease out whether or not the COP-eCOMdifference is a primary deviation or a secondary compensationof the disease. The reduced COM velocity demonstrated byour participants is in agreement with previous reports ofslower turning velocity in persons with PD (i.e., task-specificbradykinesia) [14, 25]. As discussed, our data demonstratethat persons with EPD utilize a scaled motor control strategythat limits separation of the COP and the eCOM. Thiscould be due to lack of neuromuscular control of the COM,limb and trunk position, or the result of bradykinesia orrigidity-primary deviations associated with compromisedbasal ganglia function. Alternatively, the findings may alsobe the result of secondary compensation of the disease. Forexample, this strategy could be adopted due to the inabilityto generate appropriate momentum or the presence ofneuromuscular deficits, which limit adequate muscular forceproduction [26, 28]. While we did not directly measuremuscular force, our findings are consistent with reports ofreduced lower extremity force generation in persons withPD [29]. In aggregate, the cross-sectional designs of thepreviously mentioned studies and the absence of strengthmeasures in the current study limit our ability to teaseout whether or not the shorter COP-eCOM differences areprimary deviations related to a lack of neuromuscular controlor secondary compensation of the disease. Future studies thatincorporate longitudinal designs and strength measures willultimately be required to delineate these underlying factors.

A limitation of the study is that we only examinedone walking speed, one turning direction, and one turn-ing angle. Specifically, the participants were instructed towalk at their self-selected, comfortable pace and then turnto their dominant side at the stanchions and continuewalking in the new direction. Although the instructionsfor participants to walk at their “self-selected, comfort-able pace” were instituted in order to assess participantsduring their most frequently utilized walking speeds, theseinstructions are likely to have increased the variability ofwalking speed across subjects. It is not clear, however, ifincreased walking-speed variability would also increase thevariability of our primary outcome variable, COP-eCOM,because participants may select a safe turning strategy thatpreserves COP-eCOM distance, independent of walkingspeed. Moreover, individuals will often have to modify theirmovement speed (either slowing or speeding-up) duringADLs in response to external/environmental conditions (e.g.,weather, traffic lights, ground/floor frictional characteristics,obstacles, etc.). Thus, future studies should examine theeffects of speed on postural control during turning in personswith PD, and include trials “as fast as possible”, “as safe aspossible”, and at other predetermined speeds. Additionally,in order to navigate successfully, people must turn bothright and left and negotiate a variety of turning angles(although the majority of turns experienced during ADLs


are between 76◦ and 120◦) [21]. These additional directionsand turning angles should also be examined in future studydesigns.

This study also did not examine the influence of med-ication on turning behavior and COP-eCOM. Participantsthat had pharmacological treatment were stable with nofluctuations of PD symptoms and tested while they were ontheir routine therapy. At the time of testing, none of theparticipants exhibited dyskinesia, dystonia, or other signs ofinvoluntary movement. Thus, whether or not COP-eCOMdistances would have been different had the participants notbeen on their routine therapy cannot be inferred from thecurrent investigation. In a recent report, Hong and Earhartreported that although medication significantly improvedUPDRS scores and walking velocity, it did not statisticallysignificantly alter turning performance [30]. The authorswent on, however, to report that “there was evidence for[turning] improvements particularly with respect to theamplitudes of relative rotation between segment rotationswith effect sizes ranging from 0.42 to 0.70.” They notedthat their “. . .results suggest that only certain features ofimpaired turning may be responsive to anti-Parkinson’smedication.” The participants in the Hong and Earhart studywere older and had more advanced PD than participants inthe current study—making extrapolation of their findings tothe current investigation difficult. We hypothesize, however,that medication effects on turning in persons with earlyPD will be less evident. Additional studies investigating theinfluence of medication on COP-eCOM during turning willbe needed in persons with EPD to test this hypothesis.

Despite these limitations, the results of the currentstudy provide important additional evidence that functionalimpairments can be detected even in the early stages of thedisease, when clinical signs of gait disturbance are oftenabsent [5, 31]. Taken together, these reports suggest thatidentifying the movement limitations associated with EPDrequires examination of more complex tasks that increasethe challenge to the neuromuscular system, such as turningand gait initiation. The findings also suggest that the peakCOP-eCOM distance generated during turning activitiesmay be a useful index for quantifying disease severity andintervention effectiveness. In order to determine whether thepostural control strategies during step turning are sensitive todisease severity, additional studies that examine individualsacross a broader range of disease severity will be necessary.Additionally, studies will be needed to delineate the influenceof rehabilitation interventions on postural control duringturning in persons with EPD.

5. Conclusion

Compared to HC participants, persons with EPD alteredtheir postural control strategies (shorter distance betweenthe COP and the eCOM) during the step turn. Persons withEPD appear to decrease their overall movement amplitude(i.e., COM displacement, velocity) suggesting that dynamicpostural control during turning is altered even in the earlystages of PD.

Acknowledgments

The authors gratefully acknowledge Dr. Petzinger for helpwith subject recruitment and patient evaluations and pro-viding recommendations and guidance in the developmentof the study. This study was supported by a Magistro FamilyFoundation Research Grant, the Grant-in-Aid Award atAmerican Society of Biomechanics, and the James ZumbergeResearch and Innovation Award.

References

[1] S. Colnat-Coulbois, G. C. Gauchard, L. Maillard et al.,“Bilateral subthalamic nucleus stimulation improves balancecontrol in Parkinson’s disease,” Journal of Neurology, Neuro-surgery and Psychiatry, vol. 76, no. 6, pp. 780–787, 2005.

[2] F. B. Horak, D. Dimitrova, and J. G. Nutt, “Direction-specificpostural instability in subjects with Parkinson’s disease,”Experimental Neurology, vol. 193, no. 2, pp. 504–521, 2005.

[3] I. Carpinella, P. Crenna, E. Calabrese et al., “Locomotorfunction in the early stage of Parkinson’s disease,” IEEETransactions on Neural Systems and Rehabilitation Engineering,vol. 15, no. 4, pp. 543–551, 2007.

[4] M. Ferrarin, I. Carpinella, M. Rabuffetti, E. Calabrese, P.Mazzoleni, and R. Nemni, “Locomotor disorders in patientsat early stages of Parkinson’s disease: a quantitative analysis,”Conference Proceedings IEEE Engineering in Medicine andBiology Society, vol. 1, pp. 1224–1227, 2006.

[5] N. Chastan, B. Debono, D. Maltete, and J. Weber, “Discor-dance between measured postural instability and absence ofclinical symptoms in Parkinson’s disease patients in the earlystages of the disease,” Movement Disorders, vol. 23, no. 3, pp.366–372, 2008.

[6] E. Stack and A. Ashburn, “Fall events described by people withParkinson’s disease: implications for clinical interviewing andthe research agenda,” Physiotherapy Research International, vol.4, no. 3, pp. 190–200, 1999.

[7] B. R. Bloem, Y. A. M. Grimbergen, M. Cramer, M. Willemsen,and A. H. Zwinderman, “Prospective assessment of falls inParkinson’s disease,” Journal of Neurology, vol. 248, no. 11, pp.950–958, 2001.

[8] E. Stack, K. Jupp, and A. Ashburn, “Developing methodsto evaluate how people with Parkinson’s disease turn 180◦:an activity frequently associated with falls,” Disability andRehabilitation, vol. 26, no. 8, pp. 478–484, 2004.

[9] D. Xu, L. G. Carlton, and K. S. Rosengren, “Anticipatorypostural adjustments for altering direction during walking,”Journal of Motor Behavior, vol. 36, no. 3, pp. 316–326, 2004.

[10] K. Hase and R. B. Stein, “Turning strategies during humanwalking,” Journal of Neurophysiology, vol. 81, no. 6, pp. 2914–2922, 1999.

[11] P. Crenna, I. Carpinella, M. Rabuffetti et al., “The associationbetween impaired turning and normal straight walking inParkinson’s disease,” Gait and Posture, vol. 26, no. 2, pp. 172–178, 2007.

[12] A. E. Patla, A. Adkin, and T. Ballard, “Online steering:coordination and control of body center of mass, head andbody reorientation,” Experimental Brain Research, vol. 129, no.4, pp. 629–634, 1999.

[13] M. J. D. Taylor, P. Dabnichki, and S. C. Strike, “A three-dimensional biomechanical comparison between turningstrategies during the stance phase of walking,” Human Move-ment Science, vol. 24, no. 4, pp. 558–573, 2005.


[14] E. Stack and A. Ashburn, “Dysfunctional turning in Parkin-son’s disease,” Disability and Rehabilitation, vol. 30, no. 16, pp.1222–1229, 2008.


[16] J. Spildooren, S. Vercruysse, K. Desloovere, W. Vanden-berghe, E. Kerckhofs, and A. Nieuwboer, “Freezing of gait inParkinson’s disease: the impact of dual-tasking and turning,”Movement Disorders, vol. 25, no. 15, pp. 2563–2570, 2010.

[17] A. L. Hof, M. G. J. Gazendam, and W. E. Sinke, “The conditionfor dynamic stability,” Journal of Biomechanics, vol. 38, no. 1,pp. 1–8, 2005.

[18] A. L. Hof, “The “extrapolated center of mass” concept suggestsa simple control of balance in walking,” Human MovementScience, vol. 27, no. 1, pp. 112–125, 2008.


[20] M. F. Folstein, S. E. Folstein, and P. R. McHugh, “Mini mentalstate: a practical method for grading the cognitive state ofpatients for the clinician,” Journal of Psychiatric Research, vol.12, no. 3, pp. 189–198, 1975.

[21] R. Sedgman, P. Goldie, and R. Iansek, “Development of ameasure of turning during walking,” in Proceedings of theInaugural Conference of the Faculty of Health Sciences, La TrobeUniversity, Melbourne, Australia, 1994.

[22] D. A. Winter, “Biomechanics and motor control of humanmovement,” in Kinematics, pp. 13–58, John Wiley & Sons, NewJersey, NJ, USA, 2005.

[23] C. M. O’Connor, S. K. Thorpe, M. J. O’Malley, and C. L.Vaughan, “Automatic detection of gait events using kinematicdata,” Gait and Posture, vol. 25, no. 3, pp. 469–474, 2007.

[24] M. E. Hahn and L. S. Chou, “Age-related reduction in sagittalplane center of mass motion during obstacle crossing,” Journalof Biomechanics, vol. 37, no. 6, pp. 837–844, 2004.

[25] M. E. Morris, F. Huxham, J. McGinley, K. Dodd, and R. Iansek,“The biomechanics and motor control of gait in Parkinsondisease,” Clinical Biomechanics, vol. 16, no. 6, pp. 459–470,2001.

[26] M. Martin, M. Shinberg, M. Kuchibhatla, L. Ray, J. J. Carollo,and M. L. Schenkman, “Gait initiation in community-dwelling adults with Parkinson disease: comparison with olderand younger adults without the disease,” Physical Therapy, vol.82, no. 6, pp. 566–577, 2002.

[27] T. A. Buckley, C. Pitsikoulis, and C. J. Hass, “Dynamic posturalstability during sit-to-walk transitions in Parkinson diseasepatients,” Movement Disorders, vol. 23, no. 9, pp. 1274–1280,2008.

[28] C. J. Hass, D. E. Waddell, R. P. Fleming, J. L. Juncos, andR. J. Gregor, “Gait initiation and dynamic balance controlin Parkinson’s disease,” Archives of Physical Medicine andRehabilitation, vol. 86, no. 11, pp. 2172–2176, 2005.

[29] H. Nogaki, S. Kakinuma, and M. Morimatsu, “Muscle weak-ness in Parkinson’s disease: a follow-up study,” Parkinsonismand Related Disorders, vol. 8, no. 1, pp. 57–62, 2001.

[30] M. Hong and G. M. Earhart, “Effects of medication on turningdeficits in individuals with Parkinson’s disease,” Journal ofNeurologic Physical Therapy, vol. 34, no. 1, pp. 11–16, 2010.

[31] C. Zampieri, A. Salarian, P. Carlson-Kuhta, K. Aminian, J. G.Nutt, and F. B. Horak, “The instrumented timed up and gotest: potential outcome measure for disease modifying thera-pies in Parkinson’s disease,” Journal of Neurology, Neurosurgeryand Psychiatry, vol. 81, no. 2, pp. 171–176, 2010.


Review Article

Is Freezing of Gait in Parkinson’s Disease a Result ofMultiple Gait Impairments? Implications for Treatment

Meir Plotnik,1, 2, 3 Nir Giladi,1, 4 and Jeffrey M. Hausdorff1, 5, 6

1 Laboratory for Gait and Neurodynamics, Movement Disorders Unit, Department of Neurology, Tel-Aviv Sourasky Medical Center,Tel Aviv 64239, Israel

2 Department of Physiology, Sackler Faculty of Medicine, Tel-Aviv University, 69978 Tel Aviv, Israel3 The Gonda Brain Research Center, Bar Ilan University, 52900 Ramat Gan, Israel4 Department of Neurology, Sackler Faculty of Medicine, Tel-Aviv University, 69978 Tel Aviv, Israel5 Department of Physical Therapy, Sackler Faculty of Medicine, Tel-Aviv University, 69978 Tel Aviv, Israel6 Department of Medicine, Harvard Medical School, Boston, MA 02215, USA

Correspondence should be addressed to Meir Plotnik, [email protected]

Received 1 August 2011; Accepted 26 September 2011


Copyright © 2012 Meir Plotnik et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Several gait impairments have been associated with freezing of gait (FOG) in patients with Parkinson’s disease (PD). These includedeteriorations in rhythm control, gait symmetry, bilateral coordination of gait, dynamic postural control and step scaling. Wesuggest that these seemingly independent gait features may have mutual interactions which, during certain circumstances, jointlydrive the predisposed locomotion system into a FOG episode. This new theoretical framework is illustrated by the evaluation ofthe potential relationships between the so-called “sequence effect”, that is, impairments in step scaling, and gait asymmetry justprior to FOG. We further discuss what factors influence gait control to maintain functional gait. “Triggers”, for example, such asattention shifts or trajectory transitions, may precede FOG. We propose distinct categories of interventions and describe examplesof existing work that support this idea: (a) interventions which aim to maintain a good level of locomotion control especially withrespect to aspects related to FOG; (b) those that aim at avoiding FOG “triggers”; and (c) those that merely aim to escape from FOGonce it occurs. The proposed theoretical framework sets the stage for testable hypotheses regarding the mechanisms that lead toFOG and may also lead to new treatment ideas.

1. Introduction

In a recent comprehensive review on the pathogenesis offreezing of gait (FOG), Nutt et al. describe several competinghypotheses that have been put forth to explain this episodicgait disturbance that mysteriously affects many, but not allpatients with Parkinson’s disease (PD) [1]. For example,when comparing the gait of PD patients who suffer fromfreezing (PD + FOG) with that of patients who do notsuffer from the symptom (PD-FOG), we identified severalgait properties that were abnormally altered in PD + FOGpatients, even in the interictal period [2], that is, functionalwalking periods in between freezing episodes. We suggestedthat impairments in the regulation of the gait cycle (i.e., poorcontrol of rhythmicity [3], impaired bilateral coordinationof stepping [4], and increased asymmetry [5]) operate in

the background, perhaps with executive function deficits, toset the stage for FOG that occurs in response to “triggeringevents” (e.g., turning). Some researchers highlighted theideas that impairments in dynamic postural control whilewalking [6] and in step scaling [7–10] are related to thepresence of FOG in PD. Other investigators underscore theimportance of transitions and some suggest that visual-spa-tial processing is involved in the pathogenesis of FOG [11,12]. In short, a number of different, apparently competingtheories have been put forth to explain this disabling pheno-menon.

Nutt et al. astutely point out that the extant hypothesesmay not necessarily be exclusive [1]. In this paper, we take acloser look at this idea and show how multiple gait impair-ments may take place simultaneously and lead to FOG. Morespecifically, we propose that motor control mechanisms


of two or more gait features associated with FOG mayinteract with each other and, under certain circumstances,deteriorate synergistically. Once the level of deteriorationcrosses some imaginary “redline” or critical threshold, FOGoccurs.

In the following paragraphs, we introduce the theoreticalframework underlying the hypothesis that synergism in themalfunction of the control of different gait features cancause an overall effect on gait performance that leads tofreezing episodes in patients with PD. In particular, wedescribe how two seemingly independent gait features mayboth deteriorate when challenged, thereby increasing thepropensity for FOG. We illustrate this with respect to steplength scaling and gait symmetry. Then, we discuss theclinical implications of this viewpoint.

2. Theoretical Framework: Combined Effect ofChanges in Background Levels of GaitParameters Associated with FOG Determinesthe Occurrence of FOG Episode

Figure 1 heuristically illustrates the concept that a FOGepisode occurs when the overall gait performance is nolonger sufficient to support functional gait. The overall gaitperformance is an expression of the combination of multiplecontrol mechanisms; each one addresses a different aspectof walking. In Figure 1(a), the simultaneous behavior offive gait features associated with FOG in PD are depicted:(1) bilateral coordination of gait (BCG)—the control of theantiphase left-right stepping pattern; (2) gait symmetry—the control of producing similar motor program outputsto both legs, for example, equal swing times, equal steplengths (The converse of gait symmetry, gait asymmetry,GA, is a more readily measureable and can be quantifiedby contrasting the function of one leg with that of theother); (3) step scaling—the distance covered by each step;(4) Dynamic postural control—as expressed, for example, bycenter of pressure movements; and (5) Gait rhythmicity—asexpressed, for example, by stride-to-stride variability (highervariability reflects lower rhythmicity).

According to the proposed conceptual model, the indi-vidual performance in each of these domains (a) is notconstant and varies over time and (b) in some instances,performance of a given feature may be influenced byanother gait feature. For example, during the time period10–20 (arbitrary units), all gait features associated withFOG maintain a fairly constant level and operate seeminglyindependent from each other. Similarly, when the dynamicpostural control deteriorates (at time 60–70), other gaitfeatures are not influenced.

On the other hand, other gait circumstances may yieldstronger dependency between different gait features. This isillustrated, for example, in the time periods 20–30 and 40–50. Deteriorations in gait symmetry are accompanied withdeteriorations in step length scaling, with a suggestion thatBCG may also be influenced in the 40–50 time window.

Malfunctions in gait parameters that are associated withFOG can influence the propensity for FOG not only after gait

has started, but even during the preparation for walking (e.g.,start hesitation that occurs prior to the initiation of walking).In these circumstances, the motor control system is not ableto raise a specific gait parameter to a functional level, a factwhich may also influence other gait features and result inincreased propensity for FOG (lower trace in Figure 1(a)).

Theoretically, the interrelationships between differentgait features and in particular gait features associated withFOG in PD can be described by the following analyticalformulation:

Xk(t) =n∑

i=1,i /= k

fi(t,Xi), (1)

where Xk represent any one of the gait parameters associatedwith FOG and n is the number of gait features that areassociated with FOG. Equation (1) emphasizes that eachgait parameter associated with FOG is influenced by anyof the others, and this dependency varies with time, thatis, varies with the changing “circumstances.” Furthermore,the relationship between any given pair of FOG associatedgait parameters is not identical (i.e., different functions, fi,determine the dependence). In the case of one pair, it mightbe strong, while for another pair the association may beweaker.

The thick curve in Figure 1(b) represents the compoundgait performance which is the combination (not necessarilylinear) of all individual gait features associated with FOG.Denoting this parameter by X , the following analyticalrelationship can describe the relationship between X and anyindividual Xk:

X(t) = F(X1,X2, . . . ,Xn). (2)

According to this proposed framework, as long as theoverall gait performance is maintained above a border line(i.e., a threshold), functional gait is maintained. However,once the overall gait performance deteriorates below the“threshold,” FOG occurs. As illustrated, the deterioration ofgait performance and the resultant FOG may take place as aresult of poor control of one or more individual gait featuresassociated with FOG. The duration of the FOG episode isdependent upon the ability to restore the control over thegait feature(s) that deteriorates.

3. An Example: Do the “Sequence Effect” andGait Asymmetry Converge?

The theoretical concept depicted above can be exemplified byprobing the possibility that step scaling and gait asymmetryare related to each other and to FOG. Indeed, we suggestthat insight into the pathogenesis of FOG can be gained bytaking a closer look at the “sequence effect,” one of the fiveprimary hypotheses summarized by Nutt et al. [1]. Ianseket al. first described the sequence effect and its potentialcontribution to FOG [8]. In a follow-up study that wasdesigned to experimentally control the “background” steplength, Chee et al. [7] found that when patients were cuedto walk at a markedly reduced step length, FOG became


BC

G

0 10 20 30 40 50 60 70 80

Time (a.u.)

FOG zone

Step

sca

ling

con

trol

Dyn

amic

pos

tura

lco

ntr

olG

ait

rhyt

hm

icit

yG

ait

sym

met

ry

FunctionalzoneG

ait

perf

orm

ance

(a)

(b)

Figure 1: Freezing of gait and gait features deterioration. (a) Quality of performance of gait features associated with FOG (thin lines intop 5 traces) may vary over time (hypothetical data). Similarly the level of interaction between these gait features may vary with timeand in response to different circumstances or provocations (see text). BCG—Bilateral coordination of gait. (b) The combination of theperformances of the individual gait features dictates whether FOG will occur or whether functional walking will be maintained. If theoverall performance deteriorates below a certain threshold (horizontal line), then gait freezes (FOG zone). Deterioration in the overall gaitperformance can be an expression of malfunction of single gait feature associated with FOG (see text). In some cases, the deterioration ofone gait feature can cause the deterioration of one or more gait features as portrayed in (a). FOG—Freezing of gait, a.u.—arbitrary units

more prevalent in subjects with PD + FOG. In this condition,FOG was associated with a progressive decrease in step length(“sequence effect”). In self-selected step length conditions,the PD + FOG subjects walked with a reduced step lengthas compared to PD – FOG patients. The sequence effecthypothesis posits that the progressive reduction in steplength operating on a reduced background of step lengthleads to FOG [7].

The sequence effect was nicely illustrated by Chee et al.[7]. Data from a PD + FOG subject clearly show progres-sive step length reduction (see data reconstruction in Fig-ure 2(a)). This progression is not present in the PD-FOGpatient or in a healthy control subject (c.f. original figure in

Chee et al.). At the same time, close inspection suggests thatprogressive step length reduction is also accompanied by anincreased step length asymmetry. This fact jumps out fromthe “zigzag” pattern of the step length reduction sequence(i.e., the trace repeatedly goes up and down, with the “up”obtained from one leg the and the “down” from the other;see Figure 2(a)).

Given the putative role of asymmetry in FOG [5], weconducted growth calculations of asymmetry [5] of steplength based on the data reported by Chee et al. (all six tracesshown in Figure 2 of Chee et al. were included). The PD +FOG subject who displayed the sequence effect had relativelylarge step length asymmetry (36.2%), while the PD-FOG and


0

2

4

6

8

10

12

14

16

18

20

0 5 5110 20 25 30

Step

len

gth

(cm

)

Step number

(a)

GA = 1.5% GA = 4.6% GA = 55%

One leg

Second leg

Control PD + FOG

10cm

PD − FOG

(b)

4

6

8

10

12

0 50 100

stri

de(c

m)

Step

len

gth

inth

e(i

+1

)

Asymmetry in the th stride (%) i

(c)

Figure 2: Relationship between reduced step length and asymmetry in gait. (a) Reconstruction of data presented in Figure 2 of Chee et al.[7]. The data is taken from reduced step length condition from one PD + FOG patient. (b) Similar reconstruction was performed to all 6traces presented in Chee et al. In this panel, the mean step length value of one leg (white bar) is plotted on top of the mean step length valueof the other leg (grey bar), for one reduced step length trial of PD + FOG (based on the data from (a)), a PD patient who did not experienceFOG, and a control subject (based on similar reconstruction of the original traces denoted in the original figure by “b 25%01” and “c25%01,”resp.). The asymmetry coefficient for these single gait trials is depicted above. GA values indicated in the text are the means obtained fromtwo traces from each subject. GA—Gait asymmetry as expressed by step length differences between the left and right legs. (c) Average StepLength in the (i+1)th stride as a function of step length asymmetry calculated for the preceding stride (ith). Based on data from (a).

the healthy control subjects who did not exhibit the sequenceeffect had much lower values of asymmetry (5.1% and 7.3%,resp.; see Figure 2(b) for single calculations done for singletrace only). While not definitive, this finding supports theidea that more than one gait feature may be deteriorating inassociation with FOG. Further studies are needed to addressthe possibility that in reduced step length conditions, gaitasymmetry increases among patients who experience FOGjust prior to FOG, potentially another ingredient needed toproduce the faulty state that leads to FOG.

Indeed, if we seek to develop the optimal rehabilitationprogram for FOG, it is critical to move beyond a descriptionof the phenomena and to try and identify cause and ef-fect. Examination of the data in the zigzag trace (depict-ed in Figure 2(a)) shows that across the strides, step lengthis not correlated with step length asymmetry (Spearman’sρ =−0.24, P = 0.43). There is, however, a strong inverserelationship between step length and the level of asym-metry seen in the preceding stride (Spearman’s ρ =−0.76,P = 0.005): relatively increased values of asymmetry tendto precede relatively smaller step lengths in the next stride

(Figure 2(c)). In fact, Fasano et al. [13] have recently drawnsimilar conclusions from their findings on gait freezing dur-ing treadmill walking with unbalanced subthalamic nucleusdeep brain stimulation (STN-DBS; see below): “Duringpoorly coordinated gait, information from the leg with theshorter stride length . . . might conflict with the internalcueing of the opposite leg and cause the leg with the longerstride to decrease stride length . . .. In PD patients, however,the strategy might further destabilize gait and induce avicious circle of progressively shorter step length (“sequenceeffect”), resulting in FOG. Therefore, our findings indicate apossible link between two apparently unrelated pathogenetictheories of FOG: poor interlimb coordination and the“sequence effect” [13].

While intriguing, further work is, nonetheless, needed todetermine if these findings actually reflect causality. Still, itappears that there may be more to the sequence effect thanmeets the eye and it may be an oversimplification to assumethat a sole factor is behind the mysterious phenomenonknown as FOG. This example demonstrates how multiplegait deficits, for example, asymmetry, a reduced step length,


and a further decrease in the step length, are apparently sim-ultaneously related to FOG. Acting alone, they may notalways be sufficient to cause FOG. Moreover, during walkingperiods that are not interrupted by FOG, these gait deficitsare not necessarily strongly related to each other.

Further support for this idea (i.e., that the level of inter-dependency between the two gait features is not constant)was obtained by revisiting data recently collected [14]. In astudy in which we evaluated bilateral coordination of alter-nating hand tapping, we found that stride length was notsignificantly correlated with gait asymmetry or the phasecoordination index [15] in PD patients with or without FOG(P > 0.422).

Turns may be another example of this principle. Turnsare an activity of daily living that frequently leads to FOG[16–18]. Turns also place high demands on bilateral coor-dination. In addition, during a turn, step length reductionmay be exacerbated due to the need to reduce the step lengthof the inner leg. The two effects, high demand on bilateralcoordination and reduced step length, may superimpose tocause FOG. These and additional interfering effects such asattention loading should be studied in light of the possibilitythat they work synergistically.

4. Clinical Implications: How to Address theMultifactorial Aspects of Freezing of Gait?

Treating freezing of gait in PD is very complicated and to thebest of our knowledge no “magic bullet” has yet been iden-tified. On the other hand, a conceptual approach acknow-ledging that synergy and multiple influences between gaitcontrol mechanisms have an impact on FOG may enableresearchers to generate some new thinking about treatmentopportunities as well as mechanisms. This raises an inter-esting, practical question: what should be the targets oftreatments designed to reduce FOG?

Keeping in mind the theoretical idea that the overall gaitperformance and the propensity to FOG is the product ofa combination of individual gait features associated withFOG (Figure 1(b)), we turn now to discuss what may affectthe individual gait features in a way that the compoundgait performance moves from the functional zone to theFOG zone. Figure 3(a) illustrates two instances (“triggers”)that “push” gait from the functional zone to the FOG zone(denoted by black arrows). This might reflect one of manytriggers such as dividing attention while walking, somethingthat has been associated with FOG [19–21]. When a patient’sfocus of attention is diverted from walking, the likelihood forfreezing will increase since the attention demanding task isbeing dealt with at the expense of gait control. In fact, all ofthe five gait features associated with FOG that were describedin Figure 1(a) have been shown to deteriorate when subjectswith PD perform a dual task [22–25]. It is important tonote that triggers of FOG are likely to be less effective ifthe baseline overall performance of gait is enhanced (inFigure 3(a), compare the dotted trace to solid trace). This factis supported by the observation that freezing episodes are lessfrequent during the “ON” phase of the medication cycle as

compared to the “OFF” phase [16]; in the “ON” phase, manyfeatures of gait improve, enhancing overall gait performance.

Taking these considerations into account it, seems thatinterventions for the rehabilitation/treatment of FOG shouldaddress one or more of the following aspects: (a) improvingthe overall, background gait performance, in particular gaitfeatures associated with FOG, in order to perform withinthe “envelope” of the functional gait zone; (b) improving theresponse to the occurrence of FOG provoking triggers; (c)minimizing the impact of freezing on gait regulation. Belowis a brief review of recently proposed therapeutic approachesaddressing these elements.

4.1. Improving Baseline Gait Performance to Reduce FOGPropensity. Fasano et al. [13] took advantage of the fact thatin patients with bilateral implementation of electrodes forSTN-DBS asymmetric stimulation of the subthalamic nucleican result in modulation of the symmetry and coordinationbetween legs. They examined the gait of subjects withPD who suffer from the FOG symptom and showed thatuneven stimulation (stimulation voltage decrease in thebetter functioning brain side) improved BCG as comparedto the regular prescribed stimulation (by about 60%).The frequency of FOG episodes decreased 10-fold andtheir duration decreased more than 20-fold. Additional gaitfeatures (e.g., cadence and stride length) improved as well.This study illustrates how improving baseline performance ofgait feature associated with FOG (i.e., BCG and stride length)can result in a significant reduction in FOG.

Physiotherapy interventions may also be effective inchanging usual-walking gait parameters. For example, tread-mill training can be beneficial in increasing stride length, butnot cadence (i.e., rhythmicity) [26]. However, the long-termcarry-over effects of treadmill interventions still remain to beseen.

4.2. Targeting the Triggers for FOG. From a theoretical pointof view, there are two potential types of triggers that candrive the gait control system to such poor managementthat FOG will more likely occur. The first one is attentionshifts (already mentioned above). It is, therefore, reasonableto assume that an intervention that trains the subject toimprove his/her performance in dual tasking conditions willresult in the reduction of the FOG burden. To the best of ourknowledge, only a few studies administrated dual-task (DT)based intervention in subjects with PD. Results published sofar are supportive of this notion. For example, Canning et al.[27] and Yogev-Seligmann et al. [28] found improvementsin gait speed in response to DT training which was main-tained in the retention phase. Other investigators observedimprovements in stride length [29]. If training improves DTgait speed and stride length, overall performance is likely tomove away from the FOG threshold.

The second type of FOG triggering circumstances arerelated to transitions between walking trajectory types. Thesetriggers which may shift the locomotion control to theFOG zone (e.g., the transition from straight line walking toturning) are a reasonable goal for intervention. Spatial


FOG

zon

eFu

nct

ion

alga

itzo

ne

Time axis

Gait performance OFFGait performance ON

(a)

FOG

zon

eFu

nct

ion

alga

itzo

ne

Gait performanceCued gait to reduce FOG

Time axis

(b)

Figure 3: Intervening to improve overall gait performance and to reduce the FOG burden. (a) Improving gait performance in general, forexample, by maintaining a sustained effective therapeutic effect on multiple gait features associated with FOG (recall Figure 1(a)) is a targetfor therapy that will likely reduce the FOG burden (see text). The black arrows reflect two instances where FOG might normally occur whenthe patient is OFF anti-parkinsonian medications. This might represent some diversion or divided attention that increase the likelihood ofFOG. In the ON medication condition, when the overall performance is further away from the FOG threshold (horizontal line), attentionstill has a negative effect, but it is no longer sufficient to cause FOG. In general, one way of reducing the likelihood of FOG is to move theoverall gait performance further away from this threshold. (b) Online intervention may reduce the duration of FOG episodes (see text). Key:Gait performance OFF—gait during the “off” medication periods. Gait performance ON—gait during the “on” medication periods.

circumstances that challenge one or more of the gait featuresassociated with FOG can cause malfunction in that gaitfeature that will lead to overall deterioration (i.e., “trajectorytriggers”). For example, changing trajectories from straightline walking to turning poses high demands on BCG sincegait control now produces different motor programs to theaxial (inner) and the pivotal (outer) leg; this mismatch chal-lenges coordination. In addition, the step length reductionseen in the inner leg challenges the step scaling control andmay trigger the sequence effect. Likewise, walking throughnarrow passages may lead to slowness of gait and reduction instep length since choosing a leading leg for passing throughthe passage is both an attention and coordination demandingsituation.

If effective training results in adaptation of the motor sys-tem to rapid accommodation of the post-transition gait task,then the transitioning effect (i.e., between two gait patterns)may have lesser impact on gait. In a pilot study, Hong andEarhart [30] used a rotating treadmill to extensively exposesubjects with PD who suffer from FOG to circular walking.Following this training, the two subjects who participated inthe study exhibited substantial improvement and immediatereduction in freezing episodes. The authors suggested thatafter “practice of externally cued turning, a motor patternappropriate for turning may become more automatic, facil-itating the ease of switching between straight walking andturning. This may be the mechanism of improved turningability and reduced freezing following rotating treadmilltraining” [30]. Perhaps, carefully designed gait trainingprograms for particular conditions (e.g., narrow passages)will improve the response of the patient to changing gait

patterns required spatial circumstances that would otherwiseimpact on multiple gait features and likely provoke FOG.

4.3. Assistive Device for Alleviating the FOG Symptom. Inrecent years, wearable mobility measures were used in studiesthat documented locomotion patterns in patients with PD[31]. The compound gait performance which sustainedfunctional gait or “falls” into the FOG zone (Figures 1 and3) can be a subject of quantification. This quantificationmay be achieved by the analysis of data recorded by mobilitymeasures. If such quantification can take place in real time, itmight be possible to identify gait deterioration into the FOGzone. Then, an automated response in the form of externalcue can be elicited to help the patient to restore functionalgait. This concept is heuristically illustrated in Figure 3(b).The two arrows point to the places in which gait performancecross the negative threshold into the FOG zone. If efficientautomatic detection device will identify these points andelicit external cue, then the subject utilizing the informationfrom the external cueing will restore functional gait (dashedlines departing from the solid line) more rapidly. A good“candidate” for effective external cue is the rhythmic auditorystimulation which has been proven effective in improvinggait in subjects with PD (for review Lim et al. [32]).

In a pilot study, we demonstrated the feasibility of thisstrategy [33, 34]. Within a 3-second delay, a wearable devicebased on set of accelerometers identified in real time morethan 200 FOG episodes with technical sensitivity >70% andspecificity of >80%. These promising results open the venuefor assistive devices to ameliorate the FOG symptom, ratherthan to treat it.


5. Summary

This paper addresses the problem of FOG in PD from aslightly different angle. The conceptual framework states thatmore than one gait control mechanism may be impaired inassociation with FOG and that the control of gait featuresassociated with FOG may, under certain circumstances ortriggers of FOG, interact. Paradoxically, this rather complexnature of the potential pathogenesis of FOG in PD motivatespursuing more than one scheme of intervention with severaldegrees of freedom. While curing the symptom seems out ofreach at the moment, recent findings support the promisethat sooner rather than later, the symptom will be curtailedduring the daily living of patients with PD.

Acknowledgments

The authors thank the patients and staff of the MovementDisorders Unit at the Tel-Aviv Sourasky Medical Center forinvaluable assistance. The collection of data mentioned inthis paper was supported in part by grants from the TelAviv Sourasky Medical Center Grant of Excellence, Parkin-son’s disease Foundation (PDF), the National Parkinson’sFoundation (NPF), the Israeli Ministries of Health andVeteran Affairs (Grant #3000004385), the Michael J. FoxFoundation for Parkinson’s Research Fund (MJFF) and bythe DAPHNet project, “Dynamic Analysis of PhysiologicalNetworks.” DAPHNet is a Future and Emerging Technologies(FET) project supported by the European 6th FrameworkProgram, Grant no. 018474-2.

References

[1] J. G. Nutt, B. R. Bloem, N. Giladi, M. Hallett, F. B. Horak, andA. Nieuwboer, “Freezing of gait: moving forward on a myster-ious clinical phenomenon,” The Lancet Neurology, vol. 10, no.8, pp. 734–744, 2011.

[2] M. Plotnik and J. M. Hausdorff, “The role of gait rhythmicityand bilateral coordination of stepping in the pathophysiologyof freezing of gait in Parkinson’s disease,” Movement Disorders,vol. 23, no. 2, supplement 2, pp. S444–S450, 2008.

[3] J. M. Hausdorff, J. D. Schaafsma, Y. Balash, A. L. Bartels, T.Gurevich, and N. Giladi, “Impaired regulation of stride varia-bility in Parkinson’s disease subjects with freezing of gait,” Ex-perimental Brain Research, vol. 149, no. 2, pp. 187–194, 2003.

[4] M. Plotnik, N. Giladi, and J. M. Hausdorff, “Bilateral coordi-nation of walking and freezing of gait in Parkinson’s disease,”European Journal of Neuroscience, vol. 27, no. 8, pp. 1999–2006, 2008.

[5] M. Plotnik, N. Giladi, Y. Balash, C. Peretz, and J. M. Hausdorff,“Is freezing of gait in Parkinson’s disease related to asymmetricmotor function?” Annals of Neurology, vol. 57, no. 5, pp. 656–663, 2005.

[6] J. V. Jacobs, J. G. Nutt, P. Carlson-Kuhta, M. Stephens, andF. B. Horak, “Knee trembling during freezing of gait repre-sents multiple anticipatory postural adjustments,” Experimen-tal Neurology, vol. 215, no. 2, pp. 334–341, 2009.

[7] R. Chee, A. Murphy, M. Danoudis, N. Georgiou-Karistianis,and R. Iansek, “Gait freezing in Parkinsons disease and thestride length sequence effect interaction,” Brain, vol. 132, no.8, pp. 2151–2160, 2009.

[8] R. Iansek, F. Huxham, and J. McGinley, “The sequence effectand gait festination in parkinson disease: contributors to freez-ing of gait?” Movement Disorders, vol. 21, no. 9, pp. 1419–1424,2006.

[9] A. Nieuwboer, R. Dom, W. De Weerdt, K. Desloovere, S.Fieuws, and E. Broens-Kaucsik, “Abnormalities of the spatio-temporal characteristics of Gait at the onset of freezing in Par-kinson’s disease,” Movement Disorders, vol. 16, no. 6, pp. 1066–1075, 2001.

[10] A. Nieuwboer, R. Dom, W. De Weerdt, K. Desloovere, L. Jans-sens, and V. Stijn, “Electromyographic profiles of gait prior toonset of freezing episodes in patients with Parkinson’s disease,”Brain, vol. 127, no. 7, pp. 1650–1660, 2004.

[11] D. Cowie, P. Limousin, A. Peters, and B. L. Day, “Insights intothe neural control of locomotion from walking through door-ways in Parkinson’s disease,” Neuropsychologia, vol. 48, no. 9,pp. 2750–2757, 2010.

[12] C. A. Lebold and Q. J. Almeida, “Evaluating the contributionsof dynamic flow to freezing of gait in parkinson’s disease,”Parkinson’s Disease, vol. 2010, Article ID 732508, 7 pages, 2010.

[13] A. Fasano, J. Herzog, E. Seifert et al., “Modulation of gait co-ordination by subthalamic stimulation improves freezing ofgait,” Movement Disorders, vol. 26, no. 5, pp. 844–851, 2011.

[14] M. Plotnik, T. Herman, E. Shaviv, M. Brozgol, N. Giladi, andJ. M. Hausdorff, “Impaired bilateral coordination of gait andupper extremity rhythmic movements in Parkinson’s disease:association with freezing of gait,” Movement Disorders, vol. 24,p. S341, 2009.

[15] M. Plotnik, N. Giladi, and J. M. Hausdorff, “A new measure forquantifying the bilateral coordination of human gait: effects ofaging and Parkinson’s disease,” Experimental Brain Research,vol. 181, no. 4, pp. 561–570, 2007.

[16] J. D. Schaafsma, Y. Balash, T. Gurevich, A. L. Bartels, J. M.Hausdorff, and N. Giladi, “Characterization of freezing of gaitsubtypes and the response of each to levodopa in Parkinson’sdisease,” European Journal of Neurology, vol. 10, no. 4, pp. 391–398, 2003.

[17] A. H. Snijders, M. J. Nijkrake, M. Bakker, M. Munneke, C.Wind, and B. R. Bloem, “Clinimetrics of freezing of gait,”Movement Disorders, vol. 23, no. 2, supplement 2, pp. S468–S474, 2008.

[18] J. Spildooren, S. Vercruysse, K. Desloovere, W. Vandenber-ghe, E. Kerckhofs, and A. Nieuwboer, “Freezing of gait inParkinson’s disease: the impact of dual-tasking and turning,”Movement Disorders, vol. 25, no. 15, pp. 2563–2570, 2010.

[19] M. Amboni, A. Cozzolino, K. Longo, M. Picillo, and P. Barone,“Freezing of gait and executive functions in patients with Par-kinson’s disease,” Movement Disorders, vol. 23, no. 3, pp. 395–400, 2008.

[20] R. Camicioli, B. S. Oken, G. Sexton, J. A. Kaye, and J. G. Nutt,“Verbal fluency task affects gait in Parkinson’s disease withmotor freezing,” Journal of Geriatric Psychiatry and Neurology,vol. 11, no. 4, pp. 181–185, 1998.

[21] S. L. Naismith, J. M. Shine, and S. J. G. Lewis, “The specificcontributions of set-shifting to freezing of gait in Parkinson’sdisease,” Movement Disorders, vol. 25, no. 8, pp. 1000–1004,2010.

[22] R. M. Allen, P. Carlson-Kuhta, J. V. Jacobs, J. G. Nutt, and F. B.Horak, “Effect of cognitive dual-task on balance in Parkinson’sdisease,” Movement Disorders, vol. 24, pp. S306–S307, 2009.

[23] M. Plotnik, N. Giladi, and J. M. Hausdorff, “Bilateral coordi-nation of gait and Parkinson’s disease: the effects of dual task-ing,” Journal of Neurology, Neurosurgery and Psychiatry, vol. 80,no. 3, pp. 347–350, 2009.


[24] G. Yogev, N. Giladi, C. Peretz, S. Springer, E. S. Simon, and J.M. Hausdorff, “Dual tasking, gait rhythmicity, and Parkinson’sdisease: which aspects of gait are attention demanding?” Euro-pean Journal of Neuroscience, vol. 22, no. 5, pp. 1248–1256,2005.

[25] G. Yogev, M. Plotnik, C. Peretz, N. Giladi, and J. M. Hausdorff,“Gait asymmetry in patients with Parkinson’s disease and eld-erly fallers: when does the bilateral coordination of gait re-quire attention?” Experimental Brain Research, vol. 177, no. 3,pp. 336–346, 2007.

[26] J. Mehrholz, R. Friis, J. Kugler, S. Twork, A. Storch, and M.Pohl, “Treadmill training for patients with Parkinson’s dis-ease,” Cochrane Database of Systematic Reviews, vol. 20, no. 1,Article ID CD007830, 2010.

[27] C. G. Canning, L. Ada, and E. Woodhouse, “Multiple-taskwalking training in people with mild to moderate Parkinson’sdisease: a pilot study,” Clinical Rehabilitation, vol. 22, no. 3, pp.226–233, 2008.

[28] G. Yogev-Seligmann, N. Giladi, M. Brozgol, and J. M. Haus-dorff, “A training program to improve gait while dual taskingin patients with parkinson’s disease: a pilot study,” Archives ofPhysical Medicine and Rehabilitation. In press.

[29] S. G. Brauer and M. E. Morris, “Can people with Parkinson’sdisease improve dual tasking when walking?” Gait and Posture,vol. 31, no. 2, pp. 229–233, 2010.

[30] M. Hong and G. M. Earhart, “Rotating treadmill trainingreduces freezing in Parkinson disease: preliminary observa-tions,” Parkinsonism and Related Disorders, vol. 14, no. 4, pp.359–363, 2008.

[31] S. T. Moore, H. G. MacDougall, and W. G. Ondo, “Ambulatorymonitoring of freezing of gait in Parkinson’s disease,” Journalof Neuroscience Methods, vol. 167, no. 2, pp. 340–348, 2008.

[32] I. Lim, E. van Wegen, C. de Goede et al., “Effects of externalrhythmical cueing on gait in patients with Parkinson’s disease:a systematic review,” Clinical Rehabilitation, vol. 19, no. 7, pp.695–713, 2005.

[33] M. Bachlin, M. Plotnik, D. Roggen, N. Giladi, J. M. Hausdorff,and G. Troster, “A wearable system to assist walking of parkin-sons disease patients benefits and challenges of context-trig-gered acoustic cueing,” Methods of Information in Medicine,vol. 49, no. 1, pp. 88–95, 2010.

[34] M. Bachlin, M. Plotnik, D. Roggen et al., “Wearable assistantfor Parkinsons disease patients with the freezing of gait sym-ptom,” IEEE Transactions on Information Technology in Bio-medicine, vol. 14, no. 2, Article ID 5325884, pp. 436–446, 2010.


Review Article

Posture and Locomotion Coupling: A Target for RehabilitationInterventions in Persons with Parkinson’s Disease

Marie-Laure Mille,1, 2, 3, 4 Robert A. Creath,5 Michelle G. Prettyman,5

Marjorie Johnson Hilliard,4 Katherine M. Martinez,4

Colum D. MacKinnon,4 and Mark W. Rogers5

1 UFRS STAPS, Universite du Sud Toulon-Var, La Garde 83957, France2 ISM, Aix-Marseille University, Marseille 13288, France3 UMR 6233, CNRS, Marseille 13288, France4 Department of Physical Therapy and Human Movement Sciences, Feinberg School of Medicine, Northwestern University,Chicago, IL 60611, USA

5 Department of Physical Therapy and Rehabilitation Science, University of Maryland School of Medicine,Baltimore, MD 21201, USA

Correspondence should be addressed to Mark W. Rogers, [email protected]



Copyright © 2012 Marie-Laure Mille et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

Disorders of posture, balance, and gait are debilitating motor manifestations of advancing Parkinson’s disease requiringrehabilitation intervention. These problems often reflect difficulties with coupling or sequencing posture and locomotion duringcomplex whole body movements linked with falls. Considerable progress has been made with demonstrating the effectiveness ofexercise interventions for individuals with Parkinson’s disease. However, gaps remain in the evidence base for specific interventionsand the optimal content of exercise interventions. Using a conceptual theoretical framework and experimental findings, thisperspective and review advances the viewpoint that rehabilitation interventions focused on separate or isolated components ofposture, balance, or gait may limit the effectiveness of current clinical practices. It is argued that treatment effectiveness maybe improved by directly targeting posture and locomotion coupling problems as causal factors contributing to balance and gaitdysfunction. This approach may help advance current clinical practice and improve outcomes in rehabilitation for persons withParkinson’s disease.

“. . .postural activity should be regarded as a function in its own right and not merely as a component of movement. . .”James Purdon Martin

1. Introduction

Disorders of posture, balance, and gait associated withfalls and related injuries are among the most debilitatingsymptoms of advancing Parkinson’s disease (PD). In hisseminal studies of these clinical sequelae in patients withpostencephalitic Parkinsonism entitled The Basal Gangliaand Posture published over forty years ago, the Britishneurologist James Purdon Martin documented with greatdetail the disorders of postural fixation, righting reactions,

and locomotion that similarly often accompany the progres-sion of idiopathic PD [1]. His observations on facilitatingfunctional movements—by gently rocking patients prior tochair rising or gait initiation, by placing bold transverselines on the walking surface, or through the use of visionto compensate for proprioceptive deficits—influenced thedevelopment of current rehabilitation approaches. PurdonMartin later summarized his perspectives on the integrationof posture and movement by emphasizing that “. . .posturalactivity-should be regarded as a function in its own right


and not merely as a component of movement. . .” [2]. Heconcluded that all of the conditions on which stepping andconsequently locomotion depend (e.g., antigravity supportof the body, equilibrium, propulsion) are postural in natureand that even stepping, which prevents the body from fallingforward, serves a postural function.

Although neurophysiological studies in both quadrupedsand humans have indicated that the control of posture andlocomotion is interdependent at many levels of the centralnervous system (CNS) encompassing multiple supraspinaland spinal networks [3–6], the ways by which locomotionmay be affected by prevailing postural conditions are notwell understood. This is particularly relevant to the problemsof postural instability and gait disorders that accompanyadvancing PD. Such problems are especially evident whenpatients with PD attempt to perform complex whole bodyposture and locomotion sequences during functional activ-ities such as gait initiation, sit-to-stand and other transfers,turning while standing and walking, and in ongoing gait.During such tasks, hesitation delays or “freezing” episodesare frequently observed, and the normal pattern of spatialand temporal sequencing between postural and locomotorelements of the task is either absent or disrupted [7]. Thus,the question arises as to whether or not at least some ofthe difficulties with locomotion experienced by individualswith PD are attributable to a dysfunction of the neuronalnetworks that mediate the coupling between posture andlocomotion. This issue appears to have important implica-tions for current rehabilitative interventions. For example,current physical therapy and rehabilitation interventions forposture, balance, and gait disorders in PD mainly focus onseparate aspects of the problems such as posture and balancetraining [8, 9] or gait training [10–14]. However, impairedcoupling between posture and locomotion could contributeto gait and mobility disorders, due not only to biomechanicallimitations but also to adaptive changes in neural control. Forexample, De Nunzio et al. have recently demonstrated thatalternate rhythmic vibration during quiet stance of bilateralparaspinal muscles affecting trunk posture produced acyclic transfer of the center of pressure mimicking the oneaccompanying body progression during walking [15]. Whenvibration was applied to the trunk musculature during gait,walking velocity, cadence, and stride length increased inboth patients with PD and controls [16]. In contrast, noeffects on gait were observed when leg muscles (soleus andtibialis anterior) were similarly vibrated. Since the paraspinalmuscles contralateral to the single support stance leg play arole in the stabilization of trunk posture during stance, theseresults suggest that proprioceptive feedback from posturalmuscles can be used to improve the coupling of postureand locomotion elements of the gait cycle, thus facilitatingperformance of the task in people with PD.

The purpose of this perspective and review is to presenta framework with supportive research findings to advancethe viewpoint that focusing rehabilitation interventions onindividual or isolated components of posture, balance, or gaitdisorders in persons with PD should be reevaluated. Instead,it is argued that the emphasis in intervention approachesought to be shifted towards therapeutic training programs

that directly target impairments in posture and locomotioncoupling as a causal factor contributing to balance and gaitdysfunction.

2. Conceptual and Theoretical Model

The difficulties with performing complex whole body pos-ture and locomotion sequences during functional activitiessuch as gait initiation [17–21], turning [22, 23], sit-to-stand [24, 25], and ongoing walking [16, 26] are commonlyaccompanied by timing delays in the coupling betweenpostural movements of the body segments and the goal-intended locomotion action (e.g., stepping release, stepredirection change in turning while walking, seat-off in chairrise, continuous walking). The conceptual and theoreticalframework for developing intervention approaches thattarget impairments in posture and locomotion coupling isillustrated by focusing on the initiation of gait. During gaitinitiation, an anticipatory postural adjustment (APA) phasenormally precedes and accompanies the initiation of thestepping phase [27–30]. For forward stepping, these APAsinvolve a sequence of muscle activations and changes in theground reaction forces (loading of the initial swing leg andunloading of the initial stance leg) that move the net centerof pressure beneath the feet backward and toward the initialswing limb. This motor sequence, which ends after heel off,produces the forces and moments necessary to propel thebody center of mass (COM) forward and towards the singlestance limb prior to stepping.

Compared with healthy control subjects, the medi-olateral (M-L) and anteroposterior (A-P) ground forcescharacterizing APAs in patients with PD are abnormallyprolonged in duration and reduced in amplitude with a delayin the sequencing between the beginning of the APA andstep onset [17, 18, 31, 32]. This delay may include abnormalpauses that disrupt the posture-movement coordination andmay precipitate freezing of gait (FOG). While the APA isnormally almost always present during voluntary stepping,it may often be absent in patients with PD [17, 19, 20]. Insuch cases, hesitation delays are readily observable. Thus, thenormal spatial and temporal coordination between the APAand stepping components of gait initiation is disrupted in PDin association with start hesitation and FOG.

In gait initiation, the anticipatory nature of the postural-step coordination appears to involve a role for motorprediction. A forward internal model (Figure 1) is a neuralmechanism that predicts (estimates) the future state of asystem given the current (actual) state and the sensorimotorcontrol signals [33–37]. The use of a forward model forcoordination between posture and locomotion could operatesuch that the neural circuits for initiating stepping wouldnormally be actively delayed until the APAs that generatethe weight transfer from bipedal to single leg support haveachieved single stance limb loading [38, 39]. This transitionin stance support reflects a change in the body center ofmass-base of support (COM-BOS) relationship. Thus, usinginternal and external feedback information, the forwardmodel would determine if the APAs have achieved thesufficient anticipated postural state (e.g., COM position and


Musculoskeletalsystem

Sensoryinformation

Voluntarystepping

Locomotioncontroller

External sensory feedback

Efference copiesForwardinternalmodel

Posturecontroller

Stepcommand

Posture assist

Posturecommand

+

−

+

−Predictedsensory feedback

Desired postural state:orientation and balance

COM-BOSrelationship

Figure 1: Forward internal model for posture and locomotion coupling during the initiation of gait. To initiate stepping locomotion, postureand locomotion networks are activated in parallel to generate motor commands where the posture network acts on the stepping controller.This motor output modifies the body center of mass-base of support (COM-BOS) relationship. With external posture assistance (e.g.,mechanical or sensory simulation of single stance limb loading) that enhances weight transfer to the single stance limb, an efferent copy ofthe motor commands and sensory information about the actual state of the body can be used by the CNS to modify the two commands inadvance based on an internal representation of the body and external environment (forward model). Online sensory information can alsomodulate posture and locomotion via external feedback.

motion relative to the BOS) before initiating the gait cycleand finishing the postural phase.

In Figure 1, the integrated networks for posture andlocomotion are activated in parallel [38–41] to generate aposture command for segmental orientation and balanceand a step command. These motor outputs will modify theCOM-BOS relationship. If an external mechanical or sensoryevent that assists with the APA by facilitating weight transferis applied early in the postural adjustment phase, sensoryinformation about the limb loading conditions, togetherwith an efference copy of the motor commands sent tothe forward model estimating the anticipated limb loadingconditions, can be used by the CNS to modify the twocommands in advance based on an internal representationof the body. Sensory information produced by movementcan also be used online to modulate posture and movementvia external feedback mechanisms. Conceivably, the postureassistance provided by external mechanical effects and/orsensorimotor enhancement could decrease the completiontime of the weight transfer compared with the predicted timeof completion without assistance and/or improve the fidelityof the information associated with the changes in limbloading reflecting postural state conditions during the APA.Based on a mismatch between the predicted and actual limbloading conditions determined from the forward internalmodel, the initiation release of stepping would be advancedin time and occur earlier. Reinforcement of the posture-locomotion coordination with posture-assisted locomotion

(PAL) training could lead to adaptive changes in the internalmodel for step initiation.

The mechanisms contributing to impaired gait initiationin PD are poorly understood. It has been hypothesizedthat postural instability and gait dysfunction in PD resultfrom alterations in the output of the basal ganglia to thepedunculopontine nucleus (PPN) in conjunction with theprogressive degeneration of the large cholinergic neuronsof this nucleus [42, 43]. The PPN has important inputsto regions of the mesencephalic extrapyramidal area andpontomedullary reticular formation that play a role in thepattern generation for locomotion and integration of postureand movement [44]. Alternatively, it has been proposed thatimpaired gait initiation results from dysfunction of the basalganglia and a resulting suppression or underactivity of thesupplementary motor area [45–47], a region of the frontalcortex critically involved in the planning and preparation formovement. Models of posture and movement coupling [38,40], such as the model presented in Figure 1, often emphasizethat the voluntary command to initiate movement, includingthe timing signal, must be integrated with brain stemand spinal centers that mediate the control of posture.Accordingly, the supplementary motor area may play a rolein providing feedforward information about the internalmodel to both the basal ganglia and posture and locomotioncontrol regions in the brain stem. The fact that levodopacan often improve gait initiation and locomotion in patientswith off-medication impairment [48] provides evidence that


alterations in basal ganglia output to both cortex and brainstem likely play a role in both the triggering of move-ment initiation and coupling of posture and locomotion.However, in advanced disease, posture and gait abnormalitiesoften become resistant to levodopa replacement therapy,suggesting that the progressive degeneration or dysfunctionof nondopaminergic regions of the neuraxis [49], such asthe PPN, becomes the principal pathology that mediates thedisordered coupling between posture and locomotion.

In PD, difficulties with achieving the postural prerequi-sites for stepping could contribute to gait initiation delays,“start hesitation,” and FOG. With postural assistance, theusually prolonged APA duration and reduced amplitudeaccompanying PD could be, respectively, shortened andincreased to enhance posture requirements and allow anearlier step onset time. Thus, the rationale for the PAL train-ing approach is that the expected limb loading conditionsassociated with weight transfer to the single stance limbare enhanced (e.g., achieved earlier and more effectively)compared with what would usually be expected without theassistance. If patients with PD retain the capacity to adapttheir putative internal model for stepping with PAL training,then it might be possible to remodel the timing sequence andother characteristics of posture and locomotion componentsof gait initiation.

3. Experimental Support

3.1. Postural Assistance with Weight Transfer Acutely EnhancesPosture and Locomotion Coupling and Performance during theInitiation of Stepping. A first study examined the influenceof a lateral postural assist on step initiation in patientswith PD and healthy controls [18]. Subjects performed self-paced rapid forward steps. In one condition, the APA wasassisted shortly after onset (i.e., triggered by a 5% changein loading force from baseline beneath the initial swingleg) with a lateral pull applied to the pelvis (toward theinitial stance side) by a motor-driven robotic system. Groundreaction forces and whole body kinematics were recorded tocharacterize the APA (extracted from the mediolateral centerof pressure displacement) and step characteristics (derivedfrom the first stepping leg ankle marker displacement).Overall, persons with PD (Hoehn and Yahr stage mean = 2.0)[50] tested off anti-parkinsonian medications had a longerAPA duration and longer first-step duration than controlsubjects. With the postural assistance, the APA durationfor both groups was shorter, the step onset time relativeto the APA onset was earlier, and the speed of the firststep became faster (i.e., step duration decreased while steplength did not change) for PD subjects (Figure 2). Theseimprovements in stepping performance could be related tothe influence of a sensory cue provided by the waist-pullstimuli. This possibility was assessed in a tug condition thatwas delivered in the same way as the posture assist butinvolved a displacement that was reduced to 25% of theassist waist-pull. The tug resulted in a stimulus that gave verylittle mechanical assistance with the lateral weight transferbut provided a vigorous stimulus to the pelvic area thatcould conceivably have been used as a timing cue to facilitate

stepping. No changes in performance from baseline wereobserved when a tug stimulus cue was presented (Figure 2).This ruled out that the posture assist was attributable tosensory cueing. It is also possible that stepping practice alonecould have accounted for the findings. A separate practicegroup is needed to definitively account for this possibility.However, the fact that a block of trials without mechanicalassistance or sensory cues was always presented either asthe second to last block or last block of trials and thatthese trials did not differ from the initial baseline for any ofthe measurements provides evidence that the effects of thepostural assist could not be attributed to practice alone.

These findings indicated that rapid step initiation couldbe acutely enhanced through external assistance that facil-itated weight transfer and thereby modified posture andlocomotion coupling in individuals with early stage PD whileoff of their anti-parkinsonian medication as well as in healthyolder people. In addition to the mechanical effects of therobotic assistance that contributed to passively shorteningthe APA duration and first-step onset timing, the neuralcircuits for initiating stepping could have been activelytriggered and modified in interaction with the enhancedsensory feedback providing information about the expectedor actual state conditions (e.g., center of mass positionand motion relative to base of support) associated with theevolving APA [38] (Figure 1).

Applying assistive mechanical displacement laterally atthe pelvis indirectly modifies the loading forces beneaththe feet that influence sensory inputs for posture and gaitcontrol [51]. Therefore, it is conceivable that if loadingforce information is important for timing the release ofthe gait cycle and other locomotion characteristics, thenAPA and step parameters should also be modifiable bydirectly manipulating the loading forces during gait initi-ation. Alterations in limb loading may also be importantbecause of past work demonstrating that patients withPD may show abnormalities in load receptor-mediatedproprioception during stance and gait [52]. Moreover, iflimb loading information is important for the control ofstep initiation as in ongoing gait, then healthy individualswould also be expected to demonstrate modifications instepping when limb load input is perturbed. Hence, wehave extended our waist-pull posture-assisted locomotionapproach by developing a controllable, vertical dropping-elevation perturbation system to induce changes in postureand locomotion coupling [53].

Eight patients with PD (modified Hoehn and YahrStage score 2.5 to 3.0) [50] and eight healthy controlsubjects performed rapid self-triggered step initiation withthe impending single stance limb positioned over a pneumat-ically actuated platform. All subjects had been experiencingstart hesitation or FOG. In perturbation trials, the APA waseither assisted by moving the stance limb ground supportsurface vertically downward (DROP) or resisted by movingit upward (ELEVATE), shortly after the onset of the APAphase. Overall, patients with PD demonstrated a longerAPA duration, longer time to first-step onset, and slowerstep speed than controls. In both groups, the DROP of thestance limb reinforced the intended APA kinetic changes


Parkinsonians off medication

Control subjects

Baseline Posturalassistance

Noassistance

Tug

Conditions

APA

du

rati

on (

ms)

0

100

200

300

400

500

600

700

800

900

∗∗∗†

(a)


Control subjects

0

75

150

225

300

375

450

525

600


Noassistance

Tug

Step

on

set/

APA

on

set

(ms)

Conditions

∗∗∗

(b)


Control subjects


Noassistance

Tug

500

520

540

560

580

600

620

640St

ep d

ura

tion

(m

s)

Conditions

∗∗∗

†

(c)

Figure 2: The group mean values plus 1 SD for (a) APA duration, (b) first-step onset time relative to APA onset, and (c) first-step durationin control subjects (CS: white bars) and subjects with Parkinson’s disease off medication (off: gray bars). The four experimental conditionsare initial baseline trials without postural assistance (Baseline), trials with lateral postural assistance (Postural assistance), follow-up trialswithout postural assistance (No assistance), and trials with a mechanical tug that provided no direct postural assistance (Tug). Data from[18]. †Significant differences between groups. ∗∗∗Significant difference between the postural assistance condition (ASSIST) and the others.

for lateral weight transfer (i.e., significant reduction in APAduration and increase in peak amplitude) and resulted inpositive changes in step characteristics (i.e., earlier timeto first-step onset and faster step) compared with otherconditions (Figure 3). In contrast, during ELEVATE trialsthat opposed the intended weight transfer forces, bothgroups rapidly adapted their stepping to preserve standingstability to the detriment of step characteristics by decreasingstep length and duration and increasing step height andfoot placement laterally. These findings suggest that sensoryinformation associated with limb load and/or foot pressureoccurring prior to the release of stepping modulates thespatial and temporal parameters of posture and locomotion

in interaction with a centrally generated feed-forward modeof neural control. Moreover, impaired step initiation in PDmay at least acutely be enhanced by augmenting the couplingbetween posture and locomotion through changes in limbload proprioception.

3.2. Training-Induced Changes in Postural and LocomotionCoupling and Performance during Step Initiation. From arehabilitation standpoint, it would be important to knowwhether longer-term changes in posture and locomotioncoupling are achievable with training. It is generally acknowl-edged that patients with PD can improve their motorperformance through practice training, but that they may


0

100

200

300

400

500

600

700

800

900

1000

Drop Elevate

ControlsPD

†∗∗∗

Baselines

Conditions

APA

du

rati

on (

ms)

(a)

0

1

2

3

4

5

6

7

8

9

ControlsPD

∗∗∗

Drop Elevate

Conditions

APA

am

plit

ude

(cm

)

Baselines

(b)

0

100

200

300

400

500

600

700

ControlsPD

Step

on

set/

APA

on

set

(ms)

†∗∗∗

Drop Elevate

Conditions

Baselines

(c)

ControlsPD

0

0.2

0.4

0.6

0.8

1.2

1

1.4 †∗∗∗

Drop Elevate

Conditions

Step

ave

rage

spe

ed (

m·s−

1)

Baselines

(d)

Figure 3: The group mean values (±1 SD) for patients with Parkinson’s disease (PD) on medication (gray bars) and healthy control subjects(white bars) during rapid self-paced step initiation under the different experimental conditions are presented for (a) APA duration, (b)APA amplitude, (c) step onset relative to the APA onset, and (d) first-step speed. Data from [53]. †Significant differences between groups.∗∗∗Significant difference between the postural assistance condition (DROP) and the other two.

achieve less improvement and take longer to change theirperformance than healthy adults [54, 55]. Thus, it isconceivable that posture-assisted training could be appliedto adaptively remodel the coupling between posture andlocomotion in PD. We have recently completed a feasibilityintervention study aimed at determining the effects of PALtraining using mechanosensory limb load assistance (i.e.,drop of support surface on single stance side) comparedwith sensory enhancement of weight transfer (i.e., vibrationof support surface on single stance side) on posture andlocomotion coupling and performance during step initiationin patients with PD.

Seven subjects (mean age = 72.9 years) with moderatePD (modified Hoehn and Yahr Stage score 2.5 to 3.0) [50]and on medications received baseline testing followed bytwice weekly PAL training for six weeks. For each trainingsession, the drop assist group performed 50 self-initiated

rapid stepping trials where the stance limb ground supportsurface was moved vertically downward by 1.5 cm over100 ms shortly after the onset of the APA phase (change insingle stance limb load vertical force by 5% from baselinestanding) similar to our earlier study [53].

A second group (mean age = 75.3 years) consisted ofeight subjects with moderate PD (modified Hoehn and YahrStage score 2 to 3) [50] received vibration assist trainingthrough mechanical vibration stimulus (200 Hz over 100 ms)of the single stance side support surface applied at the samerelative time point during the early APA phase as the stimulusfor the drop assist group. These PD subjects were testedon medications and followed the same testing and trainingschedule as the drop assist group.

Immediate posttesting completed after the six-weektraining phase indicated several training-associated improve-ments in kinetic APA and stepping kinematic variables. First,


0

0.1

0.2

0.3

0.4

0.5

0.6

0.7Rate of force loading beneath initial swing limb

pre post ret pre post ret

Drop groupVibration group

Error bars = std error

∗Normalized to total body weight (Newtons)

Nor

mal

ized∗

vert

ical

forc

e lo

adin

g ra

te (

1/s)

(a)

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18Peak loading force beneath initial swing limb

pre post ret pre post ret

Drop groupVibration group


∗Normalized to total body weight (Newtons)

Nor

mal

ized∗

peak

ver

tica

l for

ce

(b)

Figure 4: The group mean values plus 1 SEM for initial swing limb APA (a) rate of loading force and (b) peak loading force amplitudemeasured at baseline prior to posture assist locomotion (PAL) training (pre), immediately after training (post), and six weeks after thecompletion of training (ret) PD subjects in drop assist and vibration assist training groups. Unpublished data.

for APA characteristics (Figure 4), both the rate and peakamplitude of the loading force beneath the initial swinglimb for lateral weight transfer prior to stepping were,respectively, significantly increased, by 53% and 44% acrossboth training groups. Follow-up testing occurring six weeksafter the completion of training showed that these increaseswere retained. Second, significant group by time of testinginteractions for first-step kinematic data (Figure 5) showedthat both step speed and length were, respectively, increasedby 54% and 38% for the vibration assisted group betweenthe baseline and immediate posttest and remained greaterat retention testing. First-step height (Figure 5) was alsoincreased by 17%–25% for both groups between pretestingand both posttesting periods.

Although systematic investigation of the accuracy of thefollowing observations has yet to be addressed, two aspects ofthe approaches appear to be important for successful imple-mentation. First, the triggering of the posture enhancementstimulus should be activated by the subject’s self-initiatedpostural action, and, second, the time of delivery of the eventshould occur shortly after the onset of the posture event.This self-triggered and early event timing might enable theexternal information to be incorporated into the forwardcontrol of the posture and locomotion sequence.

4. Implementation of Posture-AssistedLocomotion Rehabilitation

4.1. Targeting Posture and Locomotion Coupling in the Reha-bilitation of People with Parkinson’s Disease. Better under-standing of posture and locomotion coupling problems has

significant relevance for physical therapy practice. To date,there has been a lack of interventions to directly addressposture and locomotion coupling problems. Interventionssuch as PAL hold promise for specifically enhancing orassisting with the posture requirements that precede andaccompany locomotion and other movements in order toimprove the spatial and temporal coupling. Ultimately, thegoal is to enhance posture and locomotion coupling toimprove performance in functional activities, foster greaterquality of life, and decrease fall risk.

Two recent reviews [56, 57] point out that whilemounting progress has been made with providing someevidence for the effectiveness of current exercise rehabil-itation approaches on balance and gait outcomes in PD,considerable gaps remain in the evidence base for specificinterventions and in identifying the optimal content ofexercise interventions. Part of the challenge in effectivelyaddressing these gaps in knowledge is in formulating con-ceptual and theoretical frameworks and models that take intoaccount the complexities or influential factors. Greater focuson the ways that the interrelationship or coupling betweenposture, balance, and locomotion elements advantage andconstrain functional performance would appear to be onesuch area where rethinking the framework for interventiondevelopment may be useful for advancing clinical practice.

4.2. Expanding the Application of the PAL Model. As men-tioned previously, the difficulties with performing complexwhole body posture and locomotion sequences in PDhave been observed for a range of different functionalactivities. For example, significant timing delays that have


0

0.2

0.4

0.6

0.8

1

1.2

1.4

Step

spe

ed (

m/s

)

Step speed for first step

pre post ret pre post retDrop groupVibration group


(a)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Step

len

gth

(m

)

Step length for first step



(b)

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.09

0.08

Step

hei

ght

(m)

Step height for first step



(c)

Figure 5: The group mean values plus 1 SEM for first-step (a) speed, (b) length, and (c) height measured at baseline prior to posture-assistedlocomotion (PAL) training (pre), immediately after training (post), and six weeks after the completion of training (ret) PD subjects in dropassist and vibration assist training groups. Unpublished data.

been identified for the sequencing between the anticipatoryforward weight transfer phase and the intended verticalascent phase of sit-to-stand [24] or between anticipatorysegmental body rotations that steer the COM prior tofoot redirection in gait turning have been observed in PDpatients [23]. These temporal disruptions are very analogousto the temporal disruption of posture and locomotioncoupling seen for gait initiation. Application of the PALapproach through mechanical and/or sensory enhancementof the early postural phase may trigger an earlier releaseof subsequent movement and possible enhancement ofoverall performance. Improvement in stepping patternsusing vibratory sensory stimulation of the trunk posturalmuscles during ongoing walking in persons with PD andhealthy controls, as demonstrated by De Nunzio et al.[16], provides a promising example of posture-locomotioncoupling applicable to intervention. Impairments in theinteraction between posture and whole body movement taskswill need further investigation to support the hypothesizedview of impaired coupling of posture and goal-intendedcomponents of action in individuals with PD.

5. Summary

In this perspective and review, we have advanced the viewpoint that approaches to rehabilitation interventions thatfocus on changing separate isolated components for posture,balance, and gait in persons with PD may have limitedeffectiveness due to the importance of posture and loco-motion coupling. Alternatively, there is neurophysiologicaland experimental support for the idea that posture and

locomotion are highly integrated components of actionthat require understanding of how these control functionsare interactively coupled. Moreover, there is evidence toindicate that individuals with PD have particular problemswith coupling or sequencing posture and locomotion dur-ing complex whole body movements that are associatedwith falls. Expanding or shifting current conceptual andtheoretical models of rehabilitation beyond posture/balanceand gait-centered intervention focuses by incorporatingposture and locomotion coupling problems as a target forrehabilitation outcomes may help to optimize and improvethe effectiveness of current clinical practice in this importantarea of rehabilitation for persons with PD.

Acknowledgments

Some of the work reported in this paper was supported by theUS National Institutes of Health Grant R21HD055386. Thecontributions of Yunhui Zhang, M.S., Tanya Simuni, M.D.,Mario Inacio, M.S., Judith Morgia, B.S., and Lisa Shulman,M.D., are gratefully acknowledged.

References

[1] J. Purdon Martin, The Basal Ganglia and Posture, PitmanMemcal Publishing Co., London, UK, 1967.

[2] J. P. Martin, “A short essay on posture and movement,” Journalof Neurology Neurosurgery and Psychiatry, vol. 40, no. 1, pp.25–29, 1977.

[3] R. Grasso, M. Zago, and F. Lacquaniti, “Interactions betweenposture and locomotion: motor patterns in humans walking


with bent posture versus erect posture,” Journal of Neurophys-iology, vol. 83, no. 1, pp. 288–300, 2000.

[4] T. Drew, S. Prentice, and B. Schepens, “Cortical and brainstemcontrol of locomotion,” Progress in Brain Research, vol. 143,pp. 251–261, 2004.

[5] S. Grillner, P. Wallen, K. Saitoh, A. Kozlov, and B. Robertson,“Neural bases of goal-directed locomotion in vertebrates-Anoverview,” Brain Research Reviews, vol. 57, no. 1, pp. 2–12,2008.

[6] T. G. Deliagina, I. N. Beloozerova, P. V. Zelenin, and G. N.Orlovsky, “Spinal and supraspinal postural networks,” BrainResearch Reviews, vol. 57, no. 1, pp. 212–221, 2008.

[7] M. E. Morris, F. E. Huxham, J. McGinley, and R. Iansek,“Gait disorders and gait rehabilitation in Parkinson’s disease,”Advances in Neurology, vol. 87, pp. 347–361, 2001.

[8] M. Schenkman, T. M. Cutson, M. Kuchibhatla et al., “Exerciseto improve spinal flexibility and function for people withParkinson’s disease: a randomized, controlled trial,” Journal ofthe American Geriatrics Society, vol. 46, no. 10, pp. 1207–1216,1998.


[10] T. C. Rubinstein, N. Giladi, and J. M. Hausdorff, “The powerof cueing to circumvent dopamine deficits: a review of physicaltherapy treatment of gait disturbances in Parkinson’s disease,”Movement Disorders, vol. 17, no. 6, pp. 1148–1160, 2002.

[11] M. Pohl, G. Rockstroh, S. Ruckriem, G. Mrass, and J.Mehrholz, “Immediate effects of speed-dependent treadmilltraining on gait parameters in early Parkinson’s disease,”Archives of Physical Medicine and Rehabilitation, vol. 84, no.12, pp. 1760–1766, 2003.

[12] S. H. J. Keus, B. R. Bloem, D. Verbaan et al., “Physiotherapyin Parkinson’s disease: utilisation and patient satisfaction,”Journal of Neurology, vol. 251, no. 6, pp. 680–687, 2004.

[13] L. E. Dibble, D. E. Nicholson, B. Shultz, B. A. MacWilliams,R. L. Marcus, and C. Moncur, “Sensory cueing effects onmaximal speed gait initiation in persons with Parkinson’sdisease and healthy elders,” Gait and Posture, vol. 19, no. 3,pp. 215–225, 2004.

[14] Y. Jiang and K. E. Norman, “Effects of visual and auditory cueson gait initiation in people with Parkinson’s disease,” ClinicalRehabilitation, vol. 20, no. 1, pp. 36–45, 2006.

[15] A. M. De Nunzio, A. Nardone, D. Picco, J. Nilsson, andM. Schieppati, “Alternate trains of postural muscle vibrationpromote cyclic body displacement in standing parkinsonianpatients,” Movement Disorders, vol. 23, no. 15, pp. 2186–2193,2008.

[16] A. M. De Nunzio, M. Grasso, A. Nardone, M. Godi, and M.Schieppati, “Alternate rhythmic vibratory stimulation of trunkmuscles affects walking cadence and velocity in Parkinson’sdisease,” Clinical Neurophysiology, vol. 121, no. 2, pp. 240–247,2010.

[17] A. Burleigh-Jacobs, F. B. Horak, J. G. Nutt, and J. A. Obeso,“Step initiation in Parkinson’s disease: influence of levodopaand external sensory triggers,” Movement Disorders, vol. 12,no. 2, pp. 206–215, 1997.

[18] M. L. Mille, M. J. Hilliard, K. M. Martinez, T. Simuni, andM. W. Rogers, “Acute effects of a lateral postural assist onvoluntary step initiation in patients with Parkinson’s disease,”Movement Disorders, vol. 22, no. 1, pp. 20–27, 2007.

[19] P. Crenna, P. Giovannelli, and I. Piccolo, “The initiation of gaitin Parkinson’s disease,” in Motor Disturbances II, A. Berardelli,R. Benecke, M. Manfredi et al., Eds., pp. 161–173, AcademicPress, London, UK, 1990.

[20] J. V. Jacobs, J. G. Nutt, P. Carlson-Kuhta, M. Stephens, and F.B. Horak, “Knee trembling during freezing of gait representsmultiple anticipatory postural adjustments,” ExperimentalNeurology, vol. 215, no. 2, pp. 334–341, 2009.

[21] M. L. Mille, M. J. Hilliard, K. M. Martinez, T. Simuni, Y.Zhang, and M. W. Rogers, “Short-term effects of posture-assisted step training on rapid step initiation in parkinson’sdisease,” Journal of Neurologic Physical Therapy, vol. 33, no. 2,pp. 88–95, 2009.

[22] M. Hong and G. M. Earhart, “Effects of medication on turningdeficits in individuals with Parkinson’s disease,” Journal ofNeurologic Physical Therapy, vol. 34, no. 1, pp. 11–16, 2010.

[23] M. K. Y. Mak, A. Patla, and C. Hui-Chan, “Sudden turn duringwalking is impaired in people with Parkinson’s disease,”Experimental Brain Research, vol. 190, no. 1, pp. 43–51, 2008.

[24] M. K. Y. Mak and C. W. Y. Hui-Chan, “The speed of sit-to-stand can be modulated in Parkinson’s disease,” ClinicalNeurophysiology, vol. 116, no. 4, pp. 780–789, 2005.

[25] M. K. Y. Mak and C. W. Y. Hui-Chan, “Switching of movementdirection is central to Parkinsonian bradykinesia in sit-to-stand,” Movement Disorders, vol. 17, no. 6, pp. 1188–1195,2002.


[27] S. Carlsoo, “The initiation of walking,” Acta Anatomica, vol.65, no. 1, pp. 1–9, 1966.

[28] R. A. Mann, J. L. Hagy, V. White, and D. Liddell, “Theinitiation of gait,” Journal of Bone and Joint Surgery Series A,vol. 61, no. 2, pp. 232–239, 1979.

[29] D. Brunt, M. J. Lafferty, A. Mckeon, B. Goode, C. Mulhausen,and P. Polk, “Invariant characteristics of gait initiation,”American Journal of Physical Medicine and Rehabilitation, vol.70, no. 4, pp. 206–212, 1991.

[30] P. Crenna and C. Frigo, “A motor programme for theinitiation of forward-oriented movements in humans,” Journalof Physiology, vol. 437, pp. 635–653, 1991.

[31] M. Vaugoyeau, F. Viallet, S. Mesure, and J. Massion, “Coor-dination of axial rotation and step execution: deficits inParkinson’s disease,” Gait and Posture, vol. 18, no. 3, pp. 150–157, 2003.

[32] N. Gantchev, F. Viallet, R. Aurenty, and J. Massion, “Impair-ment of posturo-kinetic co-ordination during initiationof forward oriented stepping movements in parkinsonianpatients,” Electroencephalography and Clinical Neurophysiol-ogy, vol. 101, no. 2, pp. 110–120, 1996.

[33] R. C. Miall, D. J. Weir, D. M. Wolpert et al., “Is the cerebelluma Smith predictor?” Journal of Motor Behavior, vol. 25, no. 3,pp. 203–216, 1993.

[34] R. C. Miall and D. M. Wolpert, “Forward models forphysiological motor control,” Neural Networks, vol. 9, no. 8,pp. 1265–1279, 1996.

[35] M. Kawato, T. Kuroda, H. Imamizu, E. Nakano, S. Miyauchi,and T. Yoshioka, “Internal forward models in the cerebellum:fMRI study on grip force and load force coupling,” Progress inBrain Research, vol. 142, pp. 171–188, 2003.

[36] P. R. Davidson and D. M. Wolpert, “Widespread access topredictive models in the motor system: a short review,” Journalof Neural Engineering, vol. 2, no. 3, pp. S313–S319, 2005.


[37] S. Yakovenko and T. Drew, “A motor cortical contribution tothe anticipatory postural adjustments that precede reaching inthe cat,” Journal of Neurophysiology, vol. 102, no. 2, pp. 853–874, 2009.

[38] J. Massion, “Movement, posture and equilibrium: interactionand coordination,” Progress in Neurobiology, vol. 38, no. 1, pp.35–56, 1992.

[39] J. Massion, A. Alexandrov, and A. Frolov, “Why and howare posture and movement coordinated?” Progress in BrainResearch, vol. 143, pp. 13–27, 2004.

[40] J. Massion, “Postural control system,” Current Opinion inNeurobiology, vol. 4, no. 6, pp. 877–887, 1994.

[41] K. Lajoie, J.-E. Andujar, K. Pearson, and T. Drew, “Neurons inarea 5 of the posterior parietal cortex in the cat contribute tointerlimb coordination during visually guided locomotion: arole in working memory,” Journal of Neurophysiology, vol. 103,no. 4, pp. 2234–2254, 2010.

[42] R. M. Zweig, W. R. Jankel, J. C. Hedreen, R. Mayeux, and D. L.Price, “The pedunculopontine nucleus in Parkinson’s disease,”Annals of Neurology, vol. 26, no. 1, pp. 41–46, 1989.

[43] P. A. Pahapill and A. M. Lozano, “The pedunculopontinenucleus and Parkinson’s disease,” Brain, vol. 123, no. 9, pp.1767–1783, 2000.

[44] B. Schepens and T. Drew, “Independent and convergent sig-nals from the pontomedullary reticular formation contributeto the control of posture and movement during reaching in thecat,” Journal of Neurophysiology, vol. 92, no. 4, pp. 2217–2238,2004.

[45] R. Rosin, H. Topka, and J. Dichgans, “Gait initiation inParkinson’s disease,” Movement Disorders, vol. 12, no. 5, pp.682–690, 1997.

[46] M. Hallett, “The intrinsic and extrinsic aspects of freezing ofgait,” Movement Disorders, vol. 23, no. 2, pp. S439–S443, 2008.

[47] J. V. Jacobs, J. S. Lou, J. A. Kraakevik, and F. B. Horak, “Thesupplementary motor area contributes to the timing of theanticipatory postural adjustment during step initiation in par-ticipants with and without Parkinson’s disease,” Neuroscience,vol. 164, no. 2, pp. 877–885, 2009.

[48] N. Giladi, “Medical treatment of freezing of gait,” MovementDisorders, vol. 23, no. 2, pp. S482–S488, 2008.

[49] H. Braak, K. Del Tredici, U. Rub, R. A. I. De Vos, E. N. H.Jansen Steur, and E. Braak, “Staging of brain pathology relatedto sporadic Parkinson’s disease,” Neurobiology of Aging, vol. 24,no. 2, pp. 197–211, 2003.

[50] M. M. Hoehn and M. D. Yahr, “Parkinsonism: onset, progres-sion and mortality,” Neurology, vol. 17, no. 5, pp. 427–442,1967.

[51] J. Duysens, F. Clarac, and H. Cruse, “Load-regulating mecha-nisms in gait and posture: comparative aspects,” PhysiologicalReviews, vol. 80, no. 1, pp. 83–133, 2000.

[52] V. Dietz and J. Duysens, “Significance of load receptor inputduring locomotion: a review,” Gait and Posture, vol. 11, no. 2,pp. 102–110, 2000.

[53] M. W. Rogers, M. J. Hilliard, K. M. Martinez, Y. Zhang, T.Simuni, and M. L. Mille, “Perturbations of ground supportalter posture and locomotion coupling during step initiationin Parkinson’s disease,” Experimental Brain Research, vol. 208,no. 4, pp. 557–567, 2011.

[54] R. Agostino, J. N. Sanes, and M. Hallett, “Motor skill learningin Parkinson’s disease,” Journal of the Neurological Sciences, vol.139, no. 2, pp. 218–226, 1996.

[55] H. I. Krebs, N. Hogan, W. Hening, S. V. Adamovich, and H.Poizner, “Procedural motor learning in parkinson’s disease,”

Experimental Brain Research, vol. 141, no. 4, pp. 425–437,2001.

[56] G. Kwakkel, C. J. T. de Goede, and E. E. H. van Wegen, “Impactof physical therapy for Parkinson’s disease: a critical review ofthe literature,” Parkinsonism and Related Disorders, vol. 13, no.3, pp. S478–S487, 2007.

[57] L. E. Dibble, O. Addison, and E. Papa, “The effects of exerciseon balance in persons with parkinson’s disease: a systematicreview across the disability spectrum,” Journal of NeurologicPhysical Therapy, vol. 33, no. 1, pp. 14–26, 2009.


Research Article

Feasibility, Safety, and Compliance in a Randomized ControlledTrial of Physical Therapy for Parkinson’s Disease

Jennifer L. McGinley,1, 2 Clarissa Martin,1 Frances E. Huxham,1

Hylton B. Menz,3 Mary Danoudis,1, 2 Anna T. Murphy,1, 2 Jennifer J. Watts,1, 4

Robert Iansek,2 and Meg E. Morris1

1 Melbourne School of Health Sciences, Physiotherapy, The University of Melbourne, Carlton, VIC 3010, Australia2 National Parkinson Foundation Center of Excellence, Clinical Research Centre for Movement Disordersand Gait and Victorian Comprehensive Parkinson’s Program, Kingston Centre, Warrigal Road, Cheltenham, VIC 3192, Australia

3 Musculoskeletal Research Centre, Faculty of Health Sciences, La Trobe University, Bundoora, VIC 3086, Australia4 Centre for Health Economics, Monash University, Building 75, Clayton, VIC 3800, Melbourne, Australia

Correspondence should be addressed to Jennifer L. McGinley, [email protected]


Academic Editor: Terry Ellis

Copyright © 2012 Jennifer L. McGinley et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

Both efficacy and clinical feasibility deserve consideration in translation of research outcomes. This study evaluated the feasibility ofrehabilitation programs within the context of a large randomized controlled trial of physical therapy. Ambulant participants withParkinson’s disease (PD) (n = 210) were randomized into three groups: (1) progressive strength training (PST); (2) movementstrategy training (MST); or (3) control (“life skills”). PST and MST included fall prevention education. Feasibility was evaluated interms of safety, retention, adherence, and compliance measures. Time to first fall during the intervention phase did not differ acrossgroups, and adverse effects were minimal. Retention was high; only eight participants withdrew during or after the interventionphase. Strong adherence (attendance > 80%) did not differ between groups (P = .435). Compliance in the therapy groups washigh. All three programs proved feasible, suggesting they may be safely implemented for people with PD in community-basedclinical practice.

1. Introduction

Physical rehabilitation in Parkinson’s disease (PD) is agrowing field of investigation. Although a number of smallrandomized controlled trials (RCT) have reported somebenefits of physical rehabilitation programs for people withPD [1–5], the outcomes of recent systematic reviews remainequivocal [6–8]. There are many excellent examples ofclinical trials, particularly those evaluating rehabilitationoutcomes (e.g., [2, 4, 9]). The existence of such richliterature highlights the importance of ensuring high levels ofadherence and compliance with therapy protocols, as well ascarefully tracking attrition and adverse responses [10]. Thismanuscript addresses that gap.

The conduct of clinical research presents challenges; trialoutcomes can be influenced by many variables related tothe rigor of research methods employed, such that evencarefully planned and well-funded clinical trials can fail

to yield high-quality data or allow results to be translatedinto practice. In addition, exercise modalities aimed atstrengthening and preventing falls may present safety risksto the potentially frail and debilitated participants enrolledin physical rehabilitation clinical trials.

Participant retention has been identified as an issue in anumber of previous randomized controlled trials of physicalrehabilitation in PD, particularly those studies involving aninactive control group [11, 12]. It has been suggested thatoffering an alternative therapy as a control intervention mayimprove retention in nonpharmacological randomized clin-ical trials [11]. Without the option of blinding participantsto group allocation within a physical rehabilitation clinicaltrial, the challenge becomes developing alternative therapyprograms that offer participants an equivalent participationexperience whilst minimizing overlap with the content of theactive intervention.


Adherence and compliance are key variables influencingthe outcome of clinical trials in PD [13]. Within the contextof a physical rehabilitation trial, it is important to establishnot only when participants attend (adherence; [1, 13]),but what activities they complete during their attendances,that is, the extent of their engagement with the program(compliance; [13]). Adherence and compliance, therefore,reflect the adequacy and appropriateness of therapy contentfor the sample and should be considered within the designof a clinical trial by the development of appropriate therapyand training protocols.

To date, few trials of physical rehabilitation programsfor PD have reported adherence and compliance data (e.g.,[14]). Fewer studies have described in detail the strategiesused or recommended to maximize adherence and com-pliance in this patient group. The purpose of this paper isto report the safety, retention, adherence, and compliancerates of a large RCT investigating the efficacy of physicalrehabilitation to reduce falls and improve mobility in peoplewith PD. In addition, strategies to improve adherence andcompliance will be described, and implications for futureresearch will be discussed.

2. Methods

2.1. Study Design. We conducted a single blind randomizedcontrolled trial to evaluate the effectiveness of two methodsof physical therapy combined with falls education to improvemobility and decrease falls in people with PD, relative to acontrol intervention [15]. Ethical approval was gained fromthe relevant Ethics Committees, and all participants providedwritten informed consent.

2.2. Participants. A convenience sample of 210 participantswith idiopathic PD was recruited between 2006 and 2009throughout Melbourne, Australia from neurologists andtherapists working in clinics and rehabilitation centers,from PD support groups and from community newspaperadvertisements. Eligible people were those who: (i) had aconfirmed diagnosis of idiopathic PD; (ii) were able towalk (Hoehn and Yahr (HY) Stages 0-IV [16]); (iii) hada Mini Mental State Examination (MMSE) score ≥24; (iv)were willing and able to attend the therapy and assessmentprogram. Exclusion criteria were other medical conditionsthat could limit or prevent exercising safely at the requiredintensity, other prior neurological conditions affecting gait,and dementia.

After screening and consent, participants were random-ized to one of three groups: progressive strength training(PST) combined with falls prevention education; movementstrategy training (MST) combined with falls preventioneducation; or a control group (life skills; LS).

2.3. Intervention. The programs were delivered by clinicalstaff employed in outpatient settings. All staff delivering theintervention completed 2.5 hours training on the therapyprotocols, conducted by the study chief investigators. Theinterventions were delivered in a once weekly two-hour

session for 8 weeks to groups of 3-4 participants. ThePST and MST interventions were delivered by a physicaltherapist, and the LS program was delivered by occupationaltherapists or social workers. An allied health assistant alsoattended sessions as required to provide general assistance. Inaddition, all participants were provided with a home exerciseprogram to be completed independently or with carer/familyassistance once a week.

The interventions are described in detail elsewhere [15,17]. To summarize, the PST program comprised sevenstrengthening exercises for core muscle groups of the lowerlimbs and trunk, in accordance with the principles of PST[18]. Exercises were progressed by adjusting the number ofsets and repetitions, by adding more weights to the vest, byincreasing the Thera-band (stretch elastic band) resistanceand by adjusting the step or chair height. Exercises wereindividually tailored and progressed, taking into accountfactors such as age, fitness level, comorbid health conditionssuch as arthritis or back pain, and self-reported exercisedifficulty according to the Borg Perceived Exertion Scale [19].The individualized home exercise program was recorded on astandardized home exercise sheet template each week by thetherapist. A booklet with photos of each exercise, a gym step,vest with weights, and Thera-band were supplied for use athome during the intervention phase.

The MST program comprised the individualized teach-ing of training strategies to enhance movement performance,improve balance and mobility, and to prevent falls, accordingto the principles outlined by Morris et al. [20–22]. Partic-ipants practised using strategies such as attention, verbal,and external cues while performing seven functional taskssuch as sit to stand, moving from chair to chair, standingand reaching, or walking and turning, either in single ordual task conditions. A booklet with photos and details ofeach exercise was provided to each participant. Exerciseswere individually tailored to the functional level of eachparticipant, and progression of each task varied accordingto need and ability. A home exercise program tailored to theindividual’s level was prescribed each week by the therapist.

Both the PST and MST groups received 10–15 minutes ofstructured falls education component at each weekly session,incorporating an overview of risk factors and strategies toprevent falls. This was based upon the content of the booklet:Don’t fall for it. Falls can be prevented!—A Guide to PreventingFalls for Older People booklet [23].

The control intervention comprised guided discussionsessions on PD-related topics such as the impact of PD onthe individual and family, fatigue management, relaxation,medication, communication, and community services. TheLS session did not include any content related to fallseducation, exercise, walking, or balance. Therapists alsosuggested activities such as reflection activities and relaxationpractice to be completed once a week at home.

2.4. Outcome Measures. All participants were tested bytrained blinded physical therapist assessors at baseline(T1), one week after the 8-week intervention (T2), andat 3 months (T3) and 12 months (T4) after intervention.The primary outcome measure was falls over 12 months


after-intervention, as detailed previously [15]. Secondaryoutcome measures included measures of mobility, activitylimitations, and quality of life.

2.5. Outcome Measures for the Intervention Phase. Interven-tion therapists recorded key details of therapy delivered aftereach session using custom designed forms. The therapyrecord for the intervention groups indicated compliance withkey therapy concepts. For the MST group, this reflectedthe individual tailoring of activities to address functionalmovement difficulties. For the PST group, the record detailedthe exercises, number of repetitions and sets, weights, andThera-band resistance level. All participants were screenedweekly at the intervention sessions for new adverse events,including new muscle soreness related to therapy. Fallswere monitored using a Falls Calendar protocol [15]. Thisrequired people to enter falls on a calendar as they occurredand to telephone a falls hotline to answer questions relatingto fall circumstances and consequences. Falls Calendars werecompleted during the intervention phase and for 12 monthsfollowing intervention.

For the purposes of this study, feasibility was adoptedas an umbrella term, encompassing the constructs of safety,retention, adherence, and compliance. Safety during theintervention phase was monitored by: (i) structured weeklyscreening by the intervention therapists for any new sorenesslasting longer than 48 hours related to therapy; (ii) recordingof adverse events that occurred during therapy; and (iii) fallrate during the intervention phase. Retention was defined inseveral ways: (i) the proportion of participants who attendedthe first post-intervention assessment; (ii) the proportion ofparticipants who returned post-intervention Falls Calendars(reflecting the primary outcome measure of the overall trial);(iii) the proportion of participants who completed all follow-up assessments compared to the number who completedbaseline assessments (note that this measure is the oppositeof “dropouts”). Adherence considered the consistency ofparticipant attendance at the intervention/control sessions.Compliance to the intervention was determined by theprogression of exercises within each of the two interventiongroups as evidenced by therapy records.

2.6. Data Analysis. Demographic data were gathered for eachgroup for variables such as age, sex, disease duration, pasthistory of falls, and comorbidities. Kaplan-Meier survivalanalysis was used to examine time to first fall during theintervention phase of the trial and compared between groupsusing Mantel-Cox log rank test. Between-group comparisonsof adherence were assessed using an independent samplesone-way Kruskal-Wallis test. Data were analysed using IBMSPSS version 19.0 (SPSS Corp, Chicago, Ill, USA) or STATA 8(Stata Corp, College Station, Tex., USA) statistical software.

3. Results

3.1. Participants. Two hundred ten participants (140 men,mean age (SD) of 67.9 (9.6), range 44–89 years) wererandomized. Participants generally had mild to moderately

severe PD, reflected by a median modified HY stage (IQR) of2.5 (2-3) and mean (SD) disease duration of 6.7 (5.6) years.Activity limitations, as measured by the Unified Parkinson’sDisease Rating Scale (UPDRS) Part II activities of dailyliving, were also mild (mean (SD); 11.6 (5.9)). One hundredsixteen participants (55%) reported having falls over theprevious 12 months, of whom 74 (64%) were repeat fallers.Freezing of gait was reported by more than half of theparticipants. Arthritis was the most commonly reportedhealth condition, present in 92 (44%) of the sample, and48 (23%) participants had a history of cancer or heartdisease. The majority of participants were taking levodopapreparations or a combination of PD medications, with 19on no PD-pharmacotherapy. One hundred fourteen (54%)participants were prescribed four or more medications, with89 (42%) taking psychotropic medication.

3.2. Delivery of Interventions. The interventions were under-taken in four different outpatient centers located in differentregions of Melbourne. Across the three years of the RCT, 8physical therapists delivered the MST, 10 physical therapistsdelivered the PST and 6 occupational therapists or socialworkers delivered the LS program. Therapist professionalexperience varied markedly from new graduate (<1 year) tohighly experienced (>30 years).

3.3. Safety. The safety of the interventions was assessed inthree ways and is reported in Table 1. Structured weeklyscreening during the intervention phase identified newsoreness lasting longer than 48 hours in 28 individuals (PSTn = 18; MST n = 10). Seven individuals reported morethan one episode of soreness (PST n = 6; MST n = 1).Typical reports included a transient increase of preexistinglow back, hip or knee pain related to osteoarthritis, resolvedby a modified program or over-the-counter medication.Fewer than one quarter of these participants attended ahealth service practitioner because of new soreness. No newsoreness was reported to persist beyond the interventionphase and require intervention.

Secondly, three incidents occurred during the actualintervention sessions. Two MST participants reported singleepisodes of dizziness with subsequent medical assessmentthat were resolved without intervention or sequelae. A singleparticipant from the PST group fell during the therapysession, with no reported injury. None of these incidentsresulted in any ongoing consequence.

The third safety evaluation examined falls in 203 partic-ipants during the intervention phase. Fifty-eight people fellduring this phase: (PST n = 10, MST n = 24, LS n = 24).Falls frequency varied markedly; 32 people fell once or twice;19 fell between 3 to 9 times; 7 fell 10 or more times. Themedian time to the first fall during the intervention phasewas 14 days in the PST group and 9 days in the MST andLS groups. The time to first fall did not differ significantlybetween groups; Log rank test (Mantel Cox), Chi square =2.08, df = 2, P = 0.353.


Table 1: Safety during the intervention phase.

PST MST LS

“New soreness > 48 hrs.”

Occasions of new soreness (n = 36) 25 11 0

Participants reporting new soreness (n = 28) 18 10 0

Incidents during therapy sessions 1 (fall; no sequelae) 2 (dizziness; no sequelae) 0

Falls during intervention phase

Number of fallers (n) 10 24 24

Falls frequency: median (IQR) 0 (0) 0 (0-1) 0 (0-1)

Falls frequency: range (n) 0–7 0–24 0–52

Median time to first fall (days) 14 9 9

Table 2: Assessments attended across the course of the trial.

Assessment PST n (% of randomized) MST n (% of randomized) LS n (% of randomized)

TI (Baseline assessment) 70 (100) 69 (100) 71 (100)

T2 (1 week after intervention phase) 69 (98.6) 68 (98.6) 59 (83.1)

T3 (3/months after intervention phase) 67 (95.7) 64 (92.8) 54 (76.1)

T4 (12 months after intervention phase) 65 (92.9) 63 (91.3) 56 (78.9)

3.4. Retention. Three aspects of retention of participantswere considered related to attendance at the three post-therapy assessments and the return of Falls Calendars. Thestudy protocol had allowed for a drop-out rate of 15% whendetermining the required sample size. Seven participants,six in the LS group and one in the MST group, withdrewprior to the intervention phase after being randomized toa group. Reasons for withdrawal included poor health (LSn = 2), a preference for the exercise group (LS n = 1),unable or no longer wanting to attend (LS n = 2, MSTn = 1), and deceased (LS n = 1). Eight participants withdrewfrom the study during or after the intervention phase anddid not return Falls Calendars during the 12 months follow-up phase. Six of these withdrew from the LS program, twodue to health reasons, one as they did not want to continue(unspecified reason), one because he felt the group was“depressing”, and two as they were not exercising or receivingfalls education. One participant withdrew from PST forhealth reasons, and one participant from the MST groupdied of unrelated causes. One hundred ninety-six (93%) ofthe participants completed the T2 assessment at the end ofthe 8-week intervention phase (PST n = 69; MST n = 68;LS/control n = 59; see Table 2).

Retention throughout the full trial period was high.One hundred ninety-five participants (93%) returned one ormore Falls Calendars during the 12-month follow-up period(PST n = 69, MST n = 67, LS n = 59). One hundred eighty-four participants (88%) provided falls data for the entire 12months (PST n = 65, MST n = 65, LS n = 54). In thefinal evaluation of retention, 775 assessments of possible 840(210 × 4 occasions) were completed (92%). Participation atthe final T4 assessment as a percentage of the total numberrandomized showed 93% of people in the PST group werereassessed, 91% of MST and 79% of participants in the LSprogram.

3.5. Adherence. Eight participants were randomized, but didnot attend any therapy sessions (PST n = 0, MST n = 2, LSn = 6). Adherence data are reported for the participants whoattended at least one intervention session (PST n = 70, MSTn = 67, LS n = 65). Ninety percent of the PST participantsattended between 6 and 8 sessions, with 3 participants (4%)attending fewer than 5 sessions. Ninety-three percent of theMST participants attended 6–8 sessions, with 2 participants(3%) attending fewer than 5 sessions. Seventy-eight percentof the LS participants attended between 6 and 8 sessions,with six participants (9%) attending fewer than 5 sessions.Participant attendance (as defined by attendance at ≥6sessions or 75%) did not differ across the three groups(independent samples Kruskal-Wallis, P = .435). The PSTgroup attended 82.5% of available sessions, the MST group90.5%, and the LS group 80.7%.

3.6. Compliance

3.6.1. Progressive Strength Training Group. A review of thetherapy records indicated that 89% of the participants wereable to complete all seven suggested exercises within the 2-hour session. The remaining 11% were able to complete sixexercises. Increasing the number of repetitions and/or setswas the most common form of progression, with 97% ofparticipants (68 of 70) progressing in this manner. Eightypercent (56 of 70) of the participants used the vest withweights during the appropriate exercises. Of these, only 5participants (9%) did not increase the weights across thecourse of the intervention. Both the step platform and Thera-band (to resist trunk extension/rotation) were used by allparticipants. Thera-band resistance was increased for 57% ofparticipants.

3.6.2. Movement Strategy Training Group. A review of theavailable therapy records (n = 64, missing data = 3)


indicated that over 86% (55/64) of the participants wereable to routinely complete six or all seven activities withinthe 2-hour period. Increasing the number of repetitions andsets was the most common form of program progression,in conjunction with increasing the difficulty of the task.Progression of the task was highly variable according to eachindividual’s task performance. For example, standing andreaching to an object in front of the participant may haveprogressed to moving the object further away, standing andplacing an object down on the ground or up on a highshelf, to moving a heavier or more cumbersome object.Similarly, walking a straight line with long steps might haveprogressed to walking with a secondary motor task, with asecondary cognitive activity, to an obstacle course; standingup from a chair may have progressed by altering the heightor compliance of the chair, or to standing up with an objectin hand or standing up and walking off.

4. Discussion

The primary RCT described in this paper investigatesthe ability of two types of physical therapy to preventfalls in community-dwelling people with PD. The currentsecondary examination of feasibility demonstrates that thesetherapy programs can be successfully implemented withinthe context of an RCT. It also suggests they are feasible andmay be safely translated to clinical practice.

4.1. Safety. As it was possible that these physical therapyprograms might present safety risks, the first aspect offeasibility considered was the safety of the two activeinterventions. Both physical therapy interventions carriedpotential risks, either inherent in their content or specificto the population being treated. Each intervention aimedto extend participants to a high level of activity andperformance, as high intensity exercise has been shown tobe achievable in people with PD [24, 25]. The possibility,thus, existed of some consequent muscle soreness and/orjoint stiffness, leading to the definition of a treatment-relatedminor adverse event as “soreness that lasted more than48 hours or required attendance at a health professional.”Falls risk was potentially increased by aspects of the MSTprogram that targeted and challenged aspects of motorperformance such as balance, reach, and stride length. ThePST program explicitly encouraged participants to work withincreasing weights and resistance, potentially risking muscle,and joint problems. Participants were primarily older people(mean age 67.9 ± 9.7 years), potentially carrying a relativelyhigh proportion of orthopedic conditions (osteoarthritis,osteoporosis) and other comorbidities [26]. Further, PDitself is strongly associated with impaired balance andincreased falls risk [1, 27]. Finally, there was the possibilitythat increased confidence, as a result of intervention, mightincrease activity or risk taking and result in further falls.Evaluation of the safety of these interventions was thereforeof key importance.

During the intervention, no adverse events with sequelaewere reported. There were no injuries during the therapy

sessions. Only 36 instances of “increased soreness > than48 hours” occurred after the 1367 sessions attended, manyin individuals with a history of back pain or osteoarthritis.A number of people were unsure whether it was theintervention or concurrent activities such as gardening orexercise that had triggered the soreness, and a visit to a healthprofessional was necessary in fewer than one quarter of theinstances reported. These results support the safety of thePST and MST programs in an older population with mild-moderate disability with a range of comorbidities.

There were no group differences between the time tofirst fall during the intervention phase. This suggestedthat neither working to improve participants’ functionalmotor performance in the MST group nor increasingtheir functional strength in the PST group may have ledthem to undertake risky behaviors and fall as a result ofoverconfidence. We conclude that both therapy programscan be safely implemented in this population.

4.1.1. Retention. A key factor in achieving meaningful resultsfrom an RCT is the retention of adequate participant num-bers. Retaining participants from any older population in aclinical trial over 12 months or more can be difficult [12, 13].The retention level in the current trial exceeded expectationsin all three measures relating to post-intervention assessmentand Falls Calendar data. Falls data for the full twelve monthswere available from 184 (88%) participants. This comparesvery favorably with returns of 78% over 6 months inpeople with PD [27] and over 75% return of monthly fallsquestionnaires in elderly fallers [28].

The other measures of retention, attendance at T2 andattendance at all 4 assessments over baseline attendance(14 months from randomization), achieved greater than 90percent retention, very similar to the 92% achieved over6 months by Tickle-Degnan et al. [29]. Two other RCTsin PD reported differing retention rates over 12-months asmeasured by attendance at assessments. Only 51 percent ofpeople with PD in one 12 month randomized controlledcrossover trial attended all three of the assessments [12],whereas results equivalent to ours were found in a muchsmaller (n = 56) study of Qigong with almost 94 percentof people returning for their 12 months assessment [30].

Differential attrition between intervention and controlgroups can affect the equivalence of the groups achievedby randomization at the outset of a trial [31, 32]. In mostcases, attrition tends to be greater in the control group[12, 30], although sometimes the intervention carries a levelof adverse effects that may cause more people to drop outof the active group [33, 34]. Whilst more participants werelost to follow up in our control group than in the therapygroups, the differences were small. The provision of a controlintervention that was similar in duration, group dynamics,and relevance to the exercise interventions appeared tooptimize retention and may have limited dropouts.

4.1.2. Adherence. Adherence, as defined by session atten-dance during the intervention phase, was also satisfactorywith over 80% of available sessions attended by each group.


Two smaller RCTs in PD (n = 68 and 116 resp.) reportedadherences of 93% [2] and between 86% and 92% over6 weeks intervention periods [2, 29]. Other reports ofadherence in the literature are either over much longerperiods [12, 14, 30], in different populations (e.g., [13, 28]),or involved physical therapy in the home setting [1, 9].The determinants of satisfactory adherence rates are likelyto be complex. There may be a degree of selection bias, aspeople who are willing to participate in research may be morelikely to adhere to a program than those who refuse. Otherfactors such as locale, professional supervision by physicaltherapists, and social interaction may be relevant [35].

4.1.3. Compliance. Evaluation of therapist and participantcompliance to the protocol interventions is important tointerpreting the key results of a trial. It also determineshow effectively the interventions can be implemented astreatments in the wider context. Despite this, it is seldomreported. In our study, over 85% of participants were able tocomplete all or nearly all of the prescribed exercises, despitea mean age of 68 years and mild to moderate signs of PD(median modified HY of 2.5). Only one other paper, toour knowledge, reports the ability of people with PD tocomply with the content of a therapy intervention to reducefalls [14]. In this small RCT (n = 48), compliance with 6months of exercise therapy performed primarily at home wasevaluated. Only 25% of participants were able to completeall prescribed exercises, with another 25% completing fewerthan half of them, possibly because motivation may bedifferent when exercising alone.

We believe compliance with therapy content wasenhanced by the booklet of photographically illustratedexercise descriptions provided to each participant in thetwo exercise groups. Enlarged photographs were also placedat exercise stations in the various therapy locations toimprove the accuracy of exercise performance, and correctperformance was further facilitated by the presence of botha physical therapist and a trained assistant. The therapyprotocol clearly directed that each participant should beworking at a hard but achievable level (modified PerceivedExertion Scale levels [19]) that should have fully engagedparticipation.

The strong compliance with content may also reflect theparticipants’ relationship with and confidence in the treatingtherapists as well as the therapists’ confidence in the trialexercise protocol. Importantly, the therapist was always thefinal judge of how the participant was to perform eachactivity and at what level, supporting their professional skilland understanding. As this level of compliance with programcontent was achieved by a number of different therapists ofwidely varying years of clinical experience, it appears that thecontent was well defined and easy to implement.

Although not formally assessed, many individuals fromour three groups volunteered that they had enjoyed theirparticipation. In part, this probably reflected the supportiverelationships and camaraderie that developed between groupmembers, reducing social isolation [36]. It was informallyobserved that group members supported each other despite

differing levels of disability. Such information sharing hasbeen reported as a desired outcome in PD [36]. These factorsare likely to have enhanced adherence and compliance in thetherapy groups.

An important aspect to designing a randomized con-trolled trial is setting up the control group. Ideally, a controlgroup should be exposed to similar duration and intensityof contact time as the intervention group, meeting the needsfor education, attention, and socialization. The results of thisstudy suggest that the LS program fulfilled these aims. Thegroup’s focus on PD specific topics [36] such as medicationmanagement, fatigue management, and communication wasa strong point, building on the camaraderie and mutualsupport provided by members to each other. Control groupscan often suffer from poor retention and adherence [12],particularly if they are simply a “wait list” group. Ourresults and others [11, 37] suggest that participant-relevanteducation helps improve group participation, particularlyif social interaction and support within the group can befostered. We conclude that guided small group discussions ontopics of relevance can be recommended as a viable controlprogram in the design of controlled clinical trials.

5. Conclusions

In the rehabilitation literature, there are few reports ofthe feasibility, safety, and adverse events associated withphysical therapy for people living with Parkinson’s disease.Our results address this gap and show that, when combinedwith a falls education program, strategy training and strengthtraining can be safely implemented in a community-basedsample of people with idiopathic PD. We also found thata life-skill social and education program was an effectivecontrol intervention that maintained interest without pro-viding the active ingredients of therapy. Protocols could beeasily followed by clinicians with varying levels of expertise,allowing for replication in future trials throughout the world.

Acknowledgment

This project has been funded by a Michael J Fox Foundation(US) Clinical Discovery Grant. H. B. Menz is currentlya National Health and Medical Research Council Fellow(Clinical Career Development Award, ID: 433049).

References

[1] A. Ashburn, L. Fazakarley, C. Ballinger, R. Pickering, L. D.McLellan, and C. Fitton, “A randomised controlled trial of ahome based exercise programme to reduce the risk of fallingamong people with Parkinson’s disease,” Journal of Neurology,Neurosurgery and Psychiatry, vol. 78, no. 7, pp. 678–684, 2007.

[2] T. Ellis, C. J. de Goede, R. G. Feldman, E. C. Wolters,G. Kwakkel, and R. C. Wagenaar, “Efficacy of a physicaltherapy program in patients with Parkinson’s disease: arandomized controlled trial,” Archives of Physical Medicine andRehabilitation, vol. 86, no. 4, pp. 626–632, 2005.

[3] M. E. Morris, R. Iansek, and B. Kirkwood, “A randomized con-trolled trial of movement strategies compared with exercise for


people with Parkinson’s disease,” Movement Disorders, vol. 24,no. 1, pp. 64–71, 2009.


[5] P. Trend, J. Kaye, H. Gage, C. Owen, and D. Wade, “Short-term effectiveness of intensive multidisciplinary rehabilitationfor people with Parkinson’s disease and their carers,” ClinicalRehabilitation, vol. 16, no. 7, pp. 717–725, 2002.

[6] H. Gage and L. Storey, “Rehabilitation for Parkinson’s disease:a systematic review of available evidence,” Clinical Rehabilita-tion, vol. 18, no. 5, pp. 463–482, 2004.


[8] M. Johnston and E. Chu, “Does attendance at a multidis-ciplinary outpatient rehabilitation program for people withParkinson’s disease produce quantitative short term or longterm improvements? A systematic review,” NeuroRehabilita-tion, vol. 26, no. 4, pp. 375–383, 2010.

[9] A. Nieuwboer, G. Kwakkel, L. Rochester et al., “Cueing train-ing in the home improves gait-related mobility in Parkinson’sdisease: the RESCUE trial,” Journal of Neurology, Neurosurgeryand Psychiatry, vol. 78, no. 2, pp. 134–140, 2007.

[10] S. H. J. Keus, M. Munneke, M. J. Nijkrake, G. Kwakkel, and B.R. Bloem, “Physical therapy in Parkinson’s disease: evolutionand future challenges,” Movement Disorders, vol. 24, no. 1, pp.1–14, 2009.


[12] D. T. Wade, H. Gage, C. Owen, P. Trend, C. Grossmith,and J. Kaye, “Multidisciplinary rehabilitation for people withParkinson’s disease: a randomised controlled study,” Journal ofNeurology Neurosurgery and Psychiatry, vol. 74, no. 2, pp. 158–162, 2003.

[13] E. V. Cyarto, W. J. Brown, and A. L. Marshall, “Retention,adherence and compliance: important considerations forhome- and group-based resistance training programs for olderadults,” Journal of Science and Medicine in Sport, vol. 9, no. 5,pp. 402–412, 2006.


[15] M. Morris et al., “Falls and mobility in Parkinson’s disease:protocol for a randomised controlled clinical trial,” BMCNeurology, vol. 11, no. 93, 2011.


[17] J. J. Watts, J. L. McGinley, F. Huxham et al., “Cost effectivenessof preventing falls and improving mobility in people withParkinson disease: protocol for an economic evaluationalongside a clinical trial,” BMC Geriatrics, vol. 8, p. 23, 2008.

[18] W. J. Kraemer, K. Adams, E. Cafarelli et al., “Progressionmodels in resistance training for healthy adults,” Medicine andScience in Sports and Exercise, vol. 34, no. 2, pp. 364–380, 2002.

[19] C. Foster, J. A. Florhaug, J. Franklin et al., “A new approachto monitoring exercise training,” Journal of Strength andConditioning Research, vol. 15, no. 1, pp. 109–115, 2001.

[20] M. E. Morris, “Movement disorders in people with Parkinsondisease: a model for physical therapy,” Physical Therapy, vol.80, no. 6, pp. 578–597, 2000.

[21] M. E. Morris, “Locomotor training in people with Parkinsondisease,” Physical Therapy, vol. 86, no. 10, pp. 1426–1435,2006.


[23] Commonwealth of Australia, Don’t Fall for It. Falls Can BePrevented!—A Guide to Preventing Falls for Older People,Commonwealth of Australia, Canberra, ACT, Australia, 2007.

[24] L. E. Dibble, T. Hale, R. L. Marcus, J. P. Gerber, and P.C. LaStayo, “The safety and feasibility of high-force eccen-tric resistance exercise in persons with Parkinson’s disease,”Archives of Physical Medicine and Rehabilitation, vol. 87, no.9, pp. 1280–1282, 2006.

[25] L. E. Dibble, T. F. Hale, R. L. Marcus, J. P. Gerber, andP. C. LaStayo, “High intensity eccentric resistance trainingdecreases bradykinesia and improves quality of life in personswith Parkinson’s disease: a preliminary study,” Parkinsonismand Related Disorders, vol. 15, no. 10, pp. 752–757, 2009.

[26] F. Prince, H. Corriveau, R. Hebert, and D. A. Winter, “Gait inthe elderly,” Gait and Posture, vol. 5, no. 2, pp. 128–135, 1997.

[27] G. K. Kerr, C. J. Worringham, M. H. Cole, P. F. Lacherez,J. M. Wood, and P. A. Silburn, “Predictors of future falls inParkinson disease,” Neurology, vol. 75, no. 2, pp. 116–124,2010.

[28] S. R. Lord, S. Castell, J. Corcoran et al., “The effect of groupexercise on physical functioning and falls in frail older peopleliving in retirement villages: a randomized, controlled trial,”Journal of the American Geriatrics Society, vol. 51, no. 12, pp.1685–1692, 2003.

[29] L. Tickle-Degnen, T. Ellis, M. H. Saint-Hilaire, C. A. Thomas,and R. C. Wagenaar, “Self-management rehabilitation andhealth-related quality of life in Parkinson’s disease: a random-ized controlled trial,” Movement Disorders, vol. 25, no. 2, pp.194–204, 2010.

[30] T. Schmitz-Hubsch, D. Pyfer, K. Kielwein, R. Fimmers,T. Klockgether, and U. Wullner, “Qigong exercise for thesymptoms of Parkinson’s disease: a randomized, controlledpilot study,” Movement Disorders, vol. 21, no. 4, pp. 543–548,2006.

[31] K. F. Schulz and D. A. Grimes, “Sample size slippages inrandomised trials: exclusions and the lost and wayward,” TheLancet, vol. 359, no. 9308, pp. 781–785, 2002.

[32] D. Sackett, W. Richardson, W. Rosenberg, and R. Haynes,Evidence Based Medicine: How to Practice and Teach EBM,Churchill Livingstone, New York, NY, USA, 1997.

[33] M. L. Pollock, J. F. Carroll, J. E. Graves et al., “Injuries andadherence to walk/jog and resistance training programs in theelderly,” Medicine and Science in Sports and Exercise, vol. 23,no. 10, pp. 1194–1200, 1991.

[34] S. F. E. Praet, E. S. J. van Rooij, A. Wijtvliet et al., “Briskwalking compared with an individualised medical fitnessprogramme for patients with type 2 diabetes: a randomisedcontrolled trial,” Diabetologia, vol. 51, no. 5, pp. 736–746,2008.

[35] M. M. Madureira, L. Takayama, A. L. Gallinaro, V. F. Caparbo,R. A. Costa, and R. M. R. Pereira, “Balance training program


is highly effective in improving functional status and reducingthe risk of falls in elderly women with osteoporosis: arandomized controlled trial,” Osteoporosis International, vol.18, no. 4, pp. 419–425, 2007.

[36] M. O’Brien, K. J. Dodd, and B. Bilney, “A qualitative analysisof a progressive resistance exercise programme for people withParkinson’s disease,” Disability and Rehabilitation, vol. 30, no.18, pp. 1350–1357, 2008.

[37] W. H. Ettinger Jr., R. Burns, S. P. Messier et al., “A randomizedtrial comparing aerobic exercise and resistance exercise witha health education program in older adults with kneeosteoarthritis: the Fitness Arthritis and Seniors Trial (FAST),”Journal of the American Medical Association, vol. 277, no. 1, pp.25–31, 1997.


Research Article

Improved Dynamic Postural Task Performancewithout Improvements in Postural Responses: The Blessingand the Curse of Dopamine Replacement

K. B. Foreman, C. Wisted, O. Addison, R. L. Marcus, P. C. LaStayo, and L. E. Dibble

Department of Physical Therapy, College of Health, University of Utah, 520 Wakara Way, Salt Lake City, UT 84108, USA

Correspondence should be addressed to K. B. Foreman, [email protected]

Received 16 June 2011; Revised 14 September 2011; Accepted 15 September 2011

Academic Editor: Alice Nieuwboer

Copyright © 2012 K. B. Foreman et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Introduction. Dopamine-replacement medications may improve mobility while not improving responses to postural challengesand could therefore increase fall risk. The purpose of this study was to measure reactive postural responses and gait-relatedmobility of patients with PD during ON and OFF medication conditions. Methods. Reactive postural responses to the PullTest and performance of the Functional Gait Assessment (FGA) were recorded from 15 persons with PD during ON and OFFmedication conditions. Results. Persons with PD demonstrated no significant difference in the reactive postural responses betweenmedication conditions but demonstrated significantly better performance on the FGA when ON medications compared to OFF.Discussion/Conclusion. Dopamine-replacement medications alone may improve gait-related mobility without improvements inreactive postural responses and therefore could result in iatrogenic increases in fall risk. Rehabilitation providers should be awareof the side effects and limitations of medication treatment and implement interventions to improve postural responses.

1. Introduction

Parkinson disease (PD) is the most prominent of the hypo-kinetic disorders [1, 2]. The cardinal features of PD are tre-mor at rest, rigidity, hypokinesia, and postural instability[3, 4]. Postural instability and falls constitute major reasonsfor the serious complications in advanced PD [5, 6]. Fallsare associated with high morbidity, mortality [7], anddiminished quality of life [8, 9]. Current estimates report thatup to 70% of those with PD fall each year, and 13% fall morethan once a week [5, 10].

The majority of persons with PD will be treated withdopamine-replacement medications and the benefits of thesemedications on overall motor function and mobility arewell established [11, 12]. However, limitations of dopaminereplacement do exist. One of these limitations is the minimaleffect of dopamine-replacement medications on posturalinstability [13–15]. Coupling the benefits of increased gait-related mobility and the limitation that postural instabilityis dopamine-resistant raises the possibility that fall risk mayincrease through increased exposure to postural challenges.

With such a high incidence of falls and the apparent dop-amine-resistant nature of postural instability, an understand-ing of the extent and character of how postural responsesand gait-related mobility respond to dopamine-replacementmedication is critical for optimal rehabilitative treatment.

Despite the apparent paradox between dopamine replac-ment effects on postural responses and gait-related mobility,to our knowledge, no studies have systematically examinedthese variables in detail. As an intial step in exploring thispostural response—mobility paradox, we sought to examinethe potential differential effect of dopamine replacementon postural instability and gait-related mobility. This studyhad the following objectives: (1) quantitatively measurethe kinematic characteristics of reactive postural responsesand gait-related mobility in persons with PD during bothON and OFF medication conditions and (2) examine thespecific components of gait-related mobility (e.g., on levelsurface, speed, with change in head position, with pivots,over obstacle, with narrow support, with eyes closed, back-wards, and steps) that were dopamine-responsive. Based onprevious research [13, 14], we hypothesized that dopamine


replacement would not improve the kinematics of reac-tive postural responses. In contrast, we hypothesized thatdopamine replacement would improve performance on gait-related mobility, but only through the improvement of spe-cific components of the Functional Gait Assessment (FGA).

2. Methods

2.1. Selection of Participants. Potential participants werea sample of convenience recruited through referral fromlocal neurologists or response to advertisement in a PDsupport group newsletter. The inclusion criteria were amedically confirmed diagnosis of idiopathic PD, a stableand neurologist-optimized medication regime that includeddopamine replacement as well as other anti-Parkinson med-ications, and the ability to independently ambulate in thecommunity with or without an assistive device. PD par-ticipants were excluded from the study if they had ahistory of medical conditions (orthopedic, cardiovascular, orotherwise) that would limit their ability to participate in thestudy procedures.

2.2. Measures. The most common research paradigms toexamine medication effects on postural instability utilizesliding or rotating force plates that induce postural sway.While having high degrees of internal validity for researchpurposes, these paradigms lack external and ecologic validitybecause the floor sliding or rotating underneath a person isnot commonly encountered in daily life. Additionally, manyof these studies limit their analysis to the components of swaywhile the base of support remains fixed omitting analysisof protective steps [11, 16]. Therefore, rather than usingmeasures that lacked ecological validity, we selected the PullTest because of its wide use in clinical neurology practice.Clinically, the Pull Test became the most widely used tool forclinical evaluation of postural instability in patients with PDwhen it was incorporated into the Unified Parkinson DiseaseRating Scale (UPDRS) [17] in 1987. However, currentresearch suggests the Pull Test in isolation is not accurate inpredicting fallers, especially in the ON medication state [5,18]. Also, the Pull Test has no formal consensus on its exactexecution and low intra- and interrater consistency [5, 19].Despite these concerns, the Pull Test is one of the only clinicalbalance test that examines reactive postural responses andprovides insight into postural reflexes without being con-founded by other aspects of mobility [7]. In order to examinepostural responses, without being corrupted by mobility,the Pull Test is performed by pulling the subject’s shouldersposteriorly inducing a protective stepping response. To ourknowledge, no studies have kinematically examined the PullTest to explore the temporal and spatial characteristics inresponse to interventions such as dopamine replacement.

Ideally, community ambulation and monitoring of fallrisk would provide direct measurement of gait-relatedmobility including step counts [20], variability of ambula-tory activity [21], episodes of instability, and falls. Althoughsome research groups have demonstrated monitoring withinlimited tasks or environments [22, 23], sustained multiday

measurement is not technologically feasible at this timeand is subject to a multitude of confounding influences[24]. Because of these concerns, we selected a clinicalmeasure that is comprised of a set of posturally challenginggait tasks that a person with PD may encounter duringcommunity mobility (the FGA [25]). Previous research hassuggested that the FGA may have greater ecological validityto postural challenges during community mobility thanthe Pull Test [26–28]. Furthermore, the FGA was selectedbecause previous research has documented its validity inpeople with Parkinson disease [18, 29], vestibular disorders[25], as well as other neurologically impaired populations[30]. The FGA was administered in a standardized locationas described in the original publication [25] and is comprisedof 10 items each worth a maximum of 3 points for a totalpossible score of 30. Higher scores are indicative of morestability during-specific balance tasks.

2.3. Procedures. Prior to testing, approval for the study wasobtained from the Institutional Review Board (IRB) at theUniversity of Utah. After recruitment, the purposes andprocedures of the study were explained and all subjectssigned an IRB approved consent form. After obtaining con-sent, demographics and disease specific variables wereobtained from each participant.

All testing was conducted at the Wellness and Rehabil-itation Clinic and the Motion Capture Core Facility at theUniversity of Utah, Department of Physical Therapy, andtook place on two separate days.

For both days of testing, the clinically defined OFFmedication condition was induced by having the participantoff their dopamine-replacement medications for at least12 hours prior to testing and is consistent with CAPITguidelines for OFF medication testing [31]. After completingOFF medication testing, participants took their medicationand rested for 1 to 1.5 hours and were retested in a clinicallydefined ON medication condition.

On the first testing day, the motor subsections of theUPDRS and FGA, during both ON and OFF medication con-ditions, were conducted by one physical therapist that hadundergone standardized training on performance of theUPDRS. Because of the significant medication effects, thetester was not blinded to medication condition. In conjunc-tion with the UPDRS testing, a modified Hoehn and Yahr(H&Y) stage [32] was assigned and a single Pull Test wasperformed and rated using the standardized scoring criteria[17]. Following completion of the UPDRS, participantsperformed the FGA.

On the second day, testing was performed in the MotionCapture Core Facility. This laboratory is equipped with aneight-camera motion analysis system (Vicon Motion Sys-tems; Oxford, UK) and two force plates (AMTI; Watertown,Mass, USA). Prior to participants’ entry into the labora-tory, a static and dynamic calibration of the system wasperformed. Individual anthropometric data were recorded.Passive reflective markers were placed on bony landmarksutilizing a standardized gait analysis marker set (Plug-In-Gait, Vicon Motion Systems; Oxford, UK). Following


subject and system preparation, participants were given anexplanation of the Pull Test prior to the execution of the testtrials [17]. Once the participant gave verbal confirmationthat they understood the test, the participant was placedinto position. The examiner, using the UPDRS testingdescription, performed the Pull Test. Participants performedfive trials in both the ON and OFF medication condition. Forall trials, kinetic and kinematic data were collected at 250 Hz.

Performance of the Pull Test was characterized usingselect spatial and temporal variables rather than just usingthe observational criteria as outlined in the UPDRS. Toaccomplish this, we segregated out 5 potential temporaland spatial contributors to abnormal Pull Test performance.These variables were chosen to specifically examine temporaland spatial constructs that have been previously shown tobe affected by PD (reaction time, movement amplitude,and movement speed) [16]. The five kinematic dependentvariables were defined as follows.

(i) Step reaction time: the time latency (in seconds[sec]) from the initial examiner induced shouldermovement until the time of initial foot movement ofthe initial stepping limb.

(ii) Step length: the distance (in centimeters [cm]) fromthe static sagital plane position of the heel marker ofthe initial stepping limb to the sagital plane positionof the heel marker at initial contact of the initialstepping limb.

(iii) Step average velocity: step length divided by step time(in cm/sec). Step time was defined as the time latencyfrom initial foot movement until the time of footcontact (in sec) of the initial stepping limb.

(iv) COM displacement: the sagital plane distance (in cm)from the initial COM position to the COM positionat time of foot contact of the initial stepping limb.

(v) COM average velocity: COM displacement divided byCOM time (in cm/sec). COM time is defined as timelatency from initial COM movement until the time offoot contact (in sec) of the initial stepping limb.

For each dependent variable, the average of the firstthree fully measured trials was used as the representativedependent variable. A fully measured trial consisted of theparticipant taking at least one step backwards to regainbalance following the Pull Test and that all markers remainedvisible during the trial.

2.4. Data Analysis. All statistical analyses were performedwith SPSS 16 for Macintosh (SPSS Inc.). Descriptive statisticswere performed for demographic variables. The independentvariable used for analysis of our primary hypotheses wasmedication condition (2 levels: ON and OFF medication).Due to the relatively small sample size and the potentialfor nonnormally distributed data, in the primary analyses,between medication condition differences were comparedusing separate nonparametric tests for dependent samples.

To examine our findings in more detail, we performedseveral post hoc means of analysis. First, between-condition

effect sizes were calculated to compare the magnitude ofeffect of the kinematic variables and the FGA. In addition,we examined the changes of the specific items on theFGA in order to gain insight into the locus of effect ofmedication on FGA performance. Differences between theON and OFF medication conditions for each FGA item werecompared using separate nonparametric tests for dependentsamples and between-condition effect sizes. A determinationof whether or not an item was dopamine-responsive wasmade by examining the statistical significance, the within-medication effect size, and the number of individuals in thesample who improved on an item when ON medication. Aconservative approach was applied to this decision in thatitems were determined to be dopamine-responsive only if3 criteria were met: (1) there was statistical significancebetween medication conditions (P < 0.005), (2) there was alarge effect size (ES > 0.70), and (3) the majority of individ-uals tested demonstrated a performance improvement withdopamine-replacement medication (>7/15).

The experiment wide level of significance was set at P <0.05. However, to control for type I error risk, the overallalpha level for the tests for differences was adjusted usinga Bonferroni correction separately within the primary andpost hoc analyses (primary analyses: 0.05/6 comparisons,therefore P < 0.008 was needed for significance on individualkinematic variables, and the overall FGA; post hoc analyses:0.05/10, therefore P < 0.005 was needed for significance onindividual FGA items).

3. Results

Fifteen persons (9 male, 6 female; mean age: 67 ± 13 years)with PD (disease duration: 7.5 ± 5.0 years) participated inthis study. Their median (range) Hoehn and Yahr ratingand mean (SD) UPDRS (motor subsection) was 2.5 (2–4)and 13.7 (6.8), respectively, while ON medication and 3.0(2.5–4) and 27.6 (7.0), respectively, while OFF medication.Furthermore, 8 of the 15 participants in this study reporteda history of falls.

3.1. Comparison of Reactive Postural Responses during ONand OFF Medication Conditions. Comparison of the reactivepostural response variables recorded from the Pull Testrevealed no significant difference between ON and OFFmedication conditions. In addition, the effect sizes fordopamine replacement for all the postural response variableswere small (0.02–0.12) (Table 1, Figures 1 and 2).

3.2. Comparison of Clinical Balance Test Performance duringON and OFF Medication Conditions. Comparison of theindex scores for the FGA revealed a significant higherscore during the ON medication condition (P ≤ 0.008).Furthermore, the effect size for dopamine replacement onthe FGA score was 1.07 (Table 1, Figure 3). In addition,post hoc examination revealed that dopamine-replacement-medication-induced improvements in FGA scores werefocused on a select group of tasks (Table 2).


Table 1: Results of PD group ON and OFF medication (Mean ± SD).

Step reactiontime (sec)

Step length(cm)

Step avg velocity(cm/sec)

COMdisplacement

(cm)

COM avg velocity(cm/sec)

Pull Test (UPDRSmotor subsection

item 30)FGA

ON meds 95%CI

0.77 ± 0.39 25.94 ± 10.33 62.45 ± 17.11 19.05 ± 6.91 19.42 ± 6.59 0.73 ± 0.46 23.67 ± 4.59∗

0.56–0.99 20.22–31.67 52.98–71.93 15.23–22.88 15.77–23.07 0.48–0.99 21.12–26.21

OFF meds 95%CI

0.75 ± 0.39 25.72 ± 11.61 63.38 ± 24.05 19.66 ± 7.46 20.20 ± 7.01 1.0 ± 0.53 18.80 ± 4.80

0.53–0.97 19.29–32.15 50.07–76.70 15.53–23.80 16.32–24.09 0.70–1.30 16.14–21.46

Effect size 0.06 0.02 .05 0.09 0.12 0.56 1.07∗P ≤ 0.008.

(a)

(b)

Figure 1: Visual representation of marker data during ON (a) andOFF (b) medication testing trials. White lines depict the equality ofstep length (Mean (SD): (a) 25.94 (10.33), (b) 25.72 (11.61)).

4. Discussion

Our clinical experience and previous reports in the literaturehave suggested that dopamine replacement may have adifferential effect on reactive postural responses comparedwith gait-related mobility. Specifically, through its reductionof bradykinesia and rigidity [33, 34], it may improve gait-related mobility. Despite these improvements, laboratorystudies of reactive and anticipatory postural tasks suggestthat postural coordination is not improved [13, 35]. There-fore, with improved gait-related mobility and deficient pos-tural coordination, some individuals may have an increasedrisk of falling. This paradox was the basis for this study.

Our results agreed with our hypotheses that dopaminereplacement does not have a significant influence on reactivepostural responses as measured by the temporal and spatialcharacteristics of the Pull Test. In addition, as hypoth-esized, dopamine-replacement medication improved gait-related mobility as measured by the overall FGA score.Further investigation of the results from the FGA indicatedthat dopamine-replacement medication improved a limitednumber of items.

Ultimately, fall events in everyday life are a product ofpostural abilities and the frequency of exposure to posturalchallenges. The research designs (ON and OFF medicationtesting as well as the measures utilized) were intendedto systematically provide an initial controlled examinationof the possibility that dopamine-replacement medicationsmay improve gait-related mobility without commensurateimprovements in reactive postural responses. As an intialstep in exploring this postural response—mobility paradox,we found that this is indeed the case. Conceivably, if sucha differential effect persisted during community mobility, itcould lead to increased fall risk and falls in the communitythrough greater exposure to balance challenges and stilldeficient postural responses. Certainly, this propositionrequires further research.

5. A Measured View of the Pull Test

The validity of the Pull Test as a predictor of falls andvalue in clinical balance examinations has been questioned[18, 36, 37]. Although our results could be seen as supportfor this view, we do not interpret our findings in this way.


Shoulder reaction time

ON meds

medsOFF

0

0.2

0.4

0.6

0.8

1

1.2

1.4(s

econ

ds)

(a)

0

5

10

15

20

25

30

35

(cen

tim

eter

s)

COM displacement

ON meds

medsOFF

(b)

0

10

20

30

40

50

60

70

80

90

100

(cen

tim

eter

s/se

con

ds)

Step velocity COM velocity

ON meds

medsOFF

(c)

Figure 2: Postural response variables.

0

5

10

15

20

25

30

FGA

inde

xsc

ore

Functional gait assessment

ON meds

medsOFF

∗

Figure 3: Clinical balance test results. ∗P ≤ 0.008.

The kinematic characteristics of the Pull Test reported in thisstudy are consistent with the hypokinetic reactive posturalresponses seen in other studies [14, 27]. Few clinical balancetests examine reactive postural responses as a componentof the motor sign of postural instability. In isolation, suchinformation provides a narrow view of potential contributorsto fall risk of persons with PD in the community. However,in conjunction with other clinical balance tests, the exami-nation of reactive postural responses may provide clinicianswith a better understanding of postural instability and fallrisk in persons with PD. In addition, concerns regardingPull Test reliability may be addressed through the use of therecently proposed Push and Release Test [37] as well as theBalance Evaluation Systems Test (BESTtest and a streamlinedversion (the Mini-BEST)) [29, 38].

6. Implications for Rehabilitation

Through the analysis of the validity indices of clinical balancetests, we previously advocated for a battery of tests [39]and environmentally valid testing [18] in the examinationof fall risk in individuals with PD. Our current findings addan additional dimension to this issue. Analysis of reactivepostural responses revealed no consistent medication effect.Examination of specific FGA items suggested that tasks withstable sensory integration demands (e.g., walking on solidground with eyes open) were more likely to be dopamine-responsive. In contrast, the dopamine-nonresponsive itemsshared the constraint of fluctuating sensory integrationdemands (e.g., gait with horizontal head turns). While thisinterpretation is speculative, such findings suggest that clini-cians should not blindly accept a composite score or specificbiomechanical outcome as an indicator of fall risk or asresponse to a rehabilitation intervention. Rather, there mustbe a critical analysis of the individual task performance inorder to understand the clinical implications of examinationfindings and the potential targets for intervention.


Table 2: FGA item analysis: items were determined to be dopamine responsive if 3 criteria were met: (1) there was statistical significancebetween medication conditions, (P < 0.005) (2) there was a large effect size (ES > 0.70), and (3) the majority of individuals testeddemonstrated a performance improvement with dopamine replacement.

FGA ItemBetween-medication

condition significance levelBetween-medicationcondition effect size

Number with positivedopamine-replacement

effect

Dopamine-responsive(Yes/No)

(1) Gait on a level surface P < 0.002 1.07 9/15 Yes

(2) Change in gait speed P < 0.004 1.03 8/15 Yes

(3) Gait with horizontal head turns P < 0.017 0.63 5/15 No

(4) Gait with sustained vertical head positions P < 0.003 0.85 7/15 No

(5) Gait and pivot turn P < 0.017 0.65 7/15 No

(6) Step over obstacle P < 0.090 0.31 2/15 No

(7) Gait with narrow base of support P < 0.003 0.90 9/15 Yes

(8) Gait with eyes closed P < 0.048 0.47 5/15 No

(9) Ambulating backwards P < 0.017 0.75 7/15 No

(10) Steps P < 0.080 0.32 2/15 No

Despite the fact that postural instability appears to bea dopamine-resistant motor sign, it does not follow that itis not amenable to change. There are few studies that haveexamined the efficacy of focused rehabilitation interventionson kinematic and kinetic outcomes [40]. In the few studiesthat have examined such outcomes, there are suggestions thatreactive postural responses or postural sway may improvewith focused training of an adequate dosage [41].

7. Limitations and Directions for Research

Despite their statistical significance, these results shouldbe interpreted with caution. Future research with largersamples is needed to gain further insight into the beneficialand potentially detrimental effects of dopamine replace-ment on postural performance and falls. Furthermore, thisstudy included only persons currently taking dopamine-replacement medications, and we did not randomize theorder of the ON and OFF medication conditions. Whilesuch a cohort may reflect persons who have progressed to amoderate disease severity, persons with mild PD (Hoehn andYahr stage 1) and severe PD (Hoehn and Yahr stage 5) didnot participate in this study. Future research should examineparticipants with these characteristics as well as persons whohave undergone surgical management of their PD (such asdeep brain stimulation). Lastly, by design, this study usedconstrained outcomes, such as the Pull Test and the FGA, asan initial test of the posture and mobility paradox. Futurestudies of postural performance and falls in persons with PDshould attempt to employ validated measures of reactive andanticipatory balance responses, clinical balance abilities, andcommunity ambulatory/fall risk monitoring as outcomes.

8. Summary and Clinical Implications

Our findings suggest that dopamine-replacement medi-cations alone may improve gait-related mobility without

commensurate improvements in reactive postural responsesand therefore could result in iatrogenic increases in fall risk.Rehabilitation providers should be aware of the limitations ofdopamine-replacement treatment and implement interven-tions intended to improve postural responses.

Conflict of Interests

The authors declare that they have no conflict of interest thatcould inappropriately influence this work.

Acknowledgement

The authors acknowledge the participants in this study aswell as the NIH for funding (Grant no. 1-R15-HD056478-01).

References

[1] J. Nolte, The Human Brain: An Introduction to Its FunctionalAnatomy, Elsevier, Philadelphia, Pa, USA, 6th edition, 2009.

[2] J. Sian, M. Gerlach, M. B. H. Youdim, and P. Riederer, “Parkin-son’s disease: a major hypokinetic basal ganglia disorder,”Journal of Neural Transmission, vol. 106, no. 5-6, pp. 443–476,1999.

[3] A. L. Hunt and K. D. Sethi, “The pull test: a history,” MovementDisorders, vol. 21, no. 7, pp. 894–899, 2006.

[4] J. Jankovic, “Parkinson’s disease: clinical features and diagno-sis,” Journal of Neurology, Neurosurgery and Psychiatry, vol. 79,no. 4, pp. 368–376, 2008.

[5] P. Valkovic, H. Brozova, K. Botzel, E. Ruzicka, and J. Benetin,“Push and release test predicts better Parkinson fallers andnonfallers than the pull test: comparison in OFF and ONmedication states,” Movement Disorders, vol. 23, no. 10, pp.1453–1457, 2008.

[6] O. H. Gerlach, A. Winogrodzka, and W. E. Weber, “Clinicalproblems in the hospitalized Parkinson’s disease patient:systematic review,” Movement Disorders, vol. 26, no. 2, pp.197–208, 2011.


[7] C. J. Hass, B. R. Bloem, and M. S. Okun, “Pushing or pullingto predict falls in Parkinson disease?” Nature Clinical PracticeNeurology, vol. 4, no. 10, pp. 530–531, 2008.

[8] K. C. Stewart, H. H. Fernandez, M. S. Okun, C. E. Jacobson,and C. J. Hass, “Distribution of motor impairment influencesquality of life in Parkinson’s disease,” Movement Disorders, vol.23, no. 10, pp. 1466–1468, 2008.

[9] T. Ellis, J. T. Cavanaugh, G. M. Earhart, M. P. Ford, K. B.Foreman, and L. E. Dibble, “Which measures of physicalfunction and motor impairment best predict quality of life inParkinson’s disease?” Parkinsonism and Related Disorders, vol.17, no. 9, pp. 693–697, 2011.

[10] B. H. Wood, J. A. Bilclough, A. Bowron, and R. W. Walker,“Incidence and prediction of falls in Parkinson’s disease:a prospective multidisciplinary study,” Journal of NeurologyNeurosurgery and Psychiatry, vol. 72, no. 6, pp. 721–725, 2002.

[11] N. I. Bohnen and R. Cham, “Postural control, gait, and dopa-mine functions in parkinsonian movement disorders,” Clinicsin Geriatric Medicine, vol. 22, no. 4, pp. 797–812, 2006.

[12] L. Rochester, K. Baker, A. Nieuwboer, and D. Burn, “Targetingdopa-sensitive and dopa-resistant gait dysfunction in Parkin-son’s disease: selective responses to internal and external cues,”Movement Disorders, vol. 26, no. 3, pp. 430–435, 2011.

[13] B. R. Bloem, D. J. Beckley, J. G. Van Dijk, A. H. Zwinderman,M. P. Remler, and R. A. C. Roos, “Influence of dopaminergicmedication on automatic postural responses and balanceimpairment in Parkinson’s disease,” Movement Disorders, vol.11, no. 5, pp. 509–521, 1996.

[14] F. B. Horak, J. Frank, and J. Nutt, “Effects of dopamine onpostural control in parkinsonian subjects: scaling, set, andtone,” Journal of Neurophysiology, vol. 75, no. 6, pp. 2380–2396, 1996.

[15] E. Tunik, A. G. Feldman, and H. Poizner, “Dopamine replace-ment therapy does not restore the ability of Parkinsonianpatients to make rapid adjustments in motor strategiesaccording to changing sensorimotor contexts,” Parkinsonismand Related Disorders, vol. 13, no. 7, pp. 425–433, 2007.

[16] I. Benatru, M. Vaugoyeau, and J. P. Azulay, “Postural disordersin Parkinson’s disease,” Neurophysiologie Clinique, vol. 38, no.6, pp. 459–465, 2008.

[17] S. Fahn and R. L. Elton, “Unified Parkinson’s disease ratingscale,” in Recent Developments in Parkinson’s Disease, S. Fahn,C. D. Marsden, and D. B. Caine, Eds., chapter 4, pp. 153–163,Mac Millan Health Care Information, Florham Park, NJ, USA,1987.

[18] K. B. Foreman, O. Addison, H. S. Kim, and L. E. Dibble, “Test-ing balance and fall risk in persons with Parkinson disease,an argument for ecologically valid testing,” Parkinsonism andRelated Disorders, vol. 17, no. 3, pp. 166–171, 2011.

[19] M. Visser, J. Marinus, B. R. Bloem, H. Kisjes, B. M. Van DenBerg, and J. J. Van Hilten, “Clinical tests for the evaluationof postural instability in patients with Parkinson’s disease,”Archives of Physical Medicine and Rehabilitation, vol. 84, no.11, pp. 1669–1674, 2003.

[20] J. T. Cavanaugh, K. L. Coleman, J. M. Gaines, L. Laing, andM. C. Morey, “Using step activity monitoring to characterizeambulatory activity in community-dwelling older adults,”Journal of the American Geriatrics Society, vol. 55, no. 1, pp.120–124, 2007.

[21] J. T. Cavanaugh, N. Kochi, and N. Stergiou, “Nonlinear anal-ysis of ambulatory activity patterns in community-dwellingolder adults,” Journals of Gerontology Series A, vol. 65, no. 2,pp. 197–203, 2010.

[22] J. M. Hausdorff, “Gait dynamics in Parkinson’s disease: com-mon and distinct behavior among stride length, gait variabil-ity, and fractal-like scaling,” Chaos, vol. 19, no. 2, Article ID026113, 2009.

[23] A. Salarian, H. Russmann, F. J. G. Vingerhoets, P. R. Burkhard,and K. Aminian, “Ambulatory monitoring of physical activi-ties in patients with Parkinson’s disease,” IEEE Transactions onBiomedical Engineering, vol. 54, no. 12, pp. 2296–2299, 2007.

[24] N. Noury, A. Fleury, P. Rumeau et al., “Fall detection—principles and methods,” Conference Proceedings—IEEE Engi-neering in Medicine and Biology Society, vol. 2007, pp. 1663–1666, 2007.

[25] D. M. Wrisley, G. F. Marchetti, D. K. Kuharsky, and S. L.Whitney, “Reliability, internal consistency, and validity ofdata obtained with the functional gait assessment,” PhysicalTherapy, vol. 84, no. 10, pp. 906–918, 2004.

[26] A. Degardin, E. Houdayer, J. L. Bourriez et al., “Deficient “sen-sory” beta synchronization in Parkinson’s disease,” ClinicalNeurophysiology, vol. 120, no. 3, pp. 636–642, 2009.

[27] J. V. Jacobs and F. B. Horak, “Abnormal proprioceptive-motorintegration contributes to hypometric postural responses ofsubjects with parkinson’s disease,” Neuroscience, vol. 141, no.2, pp. 999–1009, 2006.

[28] J. Konczak, D. M. Corcos, F. Horak et al., “Proprioceptionand motor control in parkinson’s disease,” Journal of MotorBehavior, vol. 41, no. 6, pp. 543–552, 2009.

[29] A. L. Leddy, B. E. Crowner, and G. M. Earhart, “Functionalgait assessment and balance evaluation system test: reliability,validity, sensitivity, and specificity for identifying individualswith Parkinson disease who fall,” Physical Therapy, vol. 91, no.1, pp. 102–113, 2011.

[30] H. Thieme, C. Ritschel, and C. Zange, “Reliability andvalidity of the functional gait assessment (German version)in subacute stroke patients,” Archives of Physical Medicine andRehabilitation, vol. 90, no. 9, pp. 1565–1570, 2009.

[31] J. W. Langston, H. Widner, C. G. Goetz et al., “Core assessmentprogram for intracerebral transplantations (CAPIT),” Move-ment Disorders, vol. 7, no. 1, pp. 2–13, 1992.

[32] C. G. Goetz, W. Poewe, O. Rascol et al., “Movement disordersociety task force report on the Hoehn and Yahr staging scale:status and recommendations,” Movement Disorders, vol. 19,no. 9, pp. 1020–1028, 2004.

[33] C. G. Goetz, W. Poewe, O. Rascol, and C. Sampaio, “Evidence-based medical review update: pharmacological and surgicaltreatments of Parkinson’s disease: 2001 to 2004,” MovementDisorders, vol. 20, no. 5, pp. 523–539, 2005.

[34] J. M. Miyasaki, W. Martin, O. Suchowersky, W. J. Weiner, andA. E. Lang, “Practice parameter: initiation of treatment forParkinson’s disease: an evidence-based review: report of thequality standards subcommittee of the American Academy ofNeurology,” Neurology, vol. 58, no. 1, pp. 11–17, 2002.

[35] N. Shivitz, M. M. Koop, J. Fahimi, G. Heit, and H. M. Bronte-Stewart, “Bilateral subthalamic nucleus deep brain stimulationimproves certain aspects of postural control in Parkinson’sdisease, whereas medication does not,” Movement Disorders,vol. 21, no. 8, pp. 1088–1097, 2006.


[37] J. V. Jacobs, F. B. Horak, K. Van Tran, and J. G. Nutt, “Analternative clinical postural stability test for patients withParkinson’s disease,” Journal of Neurology, vol. 253, no. 11, pp.1404–1413, 2006.


[38] F. B. Horak, D. M. Wrisley, and J. Frank, “The balance evalua-tion systems test (BESTest) to differentiate balance deficits,”Physical Therapy, vol. 89, no. 5, pp. 484–498, 2009.

[39] L. E. Dibble, J. Christensen, D. J. Ballard, and K. B. Foreman,“Diagnosis of fall risk in Parkinson disease: an analysis ofindividual and collective clinical balance test interpretation,”Physical Therapy, vol. 88, no. 3, pp. 323–332, 2008.


[41] M. Jobges, G. Heuschkel, C. Pretzel, C. Illhardt, C. Renner, andH. Hummelsheim, “Repetitive training of compensatory steps:a therapeutic approach for postural instability in Parkinson’sdisease,” Journal of Neurology, Neurosurgery and Psychiatry,vol. 75, no. 12, pp. 1682–1687, 2004.


Review Article

Upper Extremity Motor Learning amongIndividuals with Parkinson’s Disease: A Meta-AnalysisEvaluating Movement Time in Simple Tasks

K. Felix,1 K. Gain,1 E. Paiva,1 K. Whitney,1 M. E. Jenkins,2 and S. J. Spaulding3

1 Faculty of Health Sciences, The University of Western Ontario, London, ON, Canada N6A 3K72 Clinical Neurological Sciences, Schulich School of Medicine and Dentistry, The University of Western Ontario, London,ON, Canada N6A 5C1

3 Faculty of Health Sciences, School of Occupational Therapy, The University of Western Ontario, London, ON, Canada N6G 1H1

Correspondence should be addressed to S. J. Spaulding, [email protected]

Received 15 April 2011; Revised 18 July 2011; Accepted 22 September 2011


Copyright © 2012 K. Felix et al. This is an open access article distributed under the Creative Commons Attribution License, whichpermits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Motor learning has been found to occur in the rehabilitation of individuals with Parkinson’s disease (PD). Through repetitivestructured practice of motor tasks, individuals show improved performance, confirming that motor learning has probably takenplace. Although a number of studies have been completed evaluating motor learning in people with PD, the sample sizes weresmall and the improvements were variable. The purpose of this meta-analysis was to determine the ability of people with PDto learn motor tasks. Studies which measured movement time in upper extremity reaching tasks and met the inclusion criteriawere included in the analysis. Results of the meta-analysis indicated that people with PD and neurologically healthy controls bothdemonstrated motor learning, characterized by a decrease in movement time during upper extremity movements. Movement timeimprovements were greater in the control group than in individuals with PD. These results support the findings that the practiceof upper extremity reaching tasks is beneficial in reducing movement time in persons with PD and has important implications forrehabilitation.

1. Introduction

Motor learning is defined as a relatively permanent changein the ability to move associated with either practice or expe-rience [1]. In neurologically healthy adults, brain activitychanges occur in the basal ganglia during the process ofmotor learning [2]. From functional MRI studies, the keychanges include a reduction of overall brain activation anda shift from cortical to more basal ganglia activity during theconsolidation phase of learning [2, 3].

Parkinson’s disease (PD) is a neurodegenerative disorderaffecting basal ganglia functioning, characterized by fourcardinal signs; bradykinesia (slowness of movement), rigidity(stiffness), resting tremor, and postural instability. Bradyki-nesia is an inherent component of PD and affects bothmovement initiation and execution [4, 5]. Motor deficits arenot the only problem in PD. Due to the dysfunction of thebasal ganglia in PD, motor learning may also be impaired.

Acquisition and retention of movement skills are impor-tant to researchers and clinicians who are involved in reha-bilitation of individuals with PD [2, 6–8]. Nieuwboer et al.(2009) [6] reviewed 11 studies that evaluated acquisition andretention in a broad range of tasks. The studies suggest thatoverall, acquisition does occur in people with PD, but per-formance on the task during acquisition is typically impairedrelative to controls. Nieuwboer et al.’s [6] review also suggeststhat long-term retention of new skills is impaired in individ-uals who have striatal problems, particularly in people withPD.

Although a number of studies have examined acquisitionand retention of tasks in PD, the sample sizes have beensmall and heterogeneous, and the experimental tasks andoutcomes used have varied widely. For example, kinematicvariables, including distance (or displacement, which isdistance with a specific direction), speed (or velocity, whichis speed with a direction), and acceleration, have been used


to measure motor learning both in the upper and the lowerextremities in individuals with PD [9, 10]. Other movementparameters that have been measured include time, force,accuracy of movement to a target, coordination of more thanone joint segment of the limb, sequencing of movement [9],interlimb function [11], and the ability to switch motor tasks[12]. Any of these measurements can provide researcherswith valuable information about motor learning abilities inindividuals with PD.

Regardless of the design features of each study, practiceof the experimental task is integral to any of the research par-adigms. While some researchers have suggested that peoplewith PD do improve with practice, but not to the same levelor as well as do control subjects [13–15], others have sug-gested that people with PD were able to benefit from short-term, but not long-term practice [16]. Sequence learning(learning of movements in a set sequence) has been shown totake more time and to be related to the stage of disease [13].

Given the apparent heterogeneity of methodologies andparticipant samples, it is not surprising that there is disagree-ment on the extent and duration of skill acquisition in per-sons with PD. Such disagreement makes it difficult to drawfirm conclusions and provide therapeutic recommendationsto clinicians. To date, there have been systematic reviews, butno meta-analyses pooling or combining the existing data onacquisition and retention of skills in individuals with PD thatmay provide insight into the consistent effects of motor taskpractice.

By focusing only on upper extremity and on movementtime during practice of upper extremity reaching tasks, wewere able to find a sufficient body of literature to analyzeusing a meta-analysis paradigm. The purpose of this study,therefore, was to determine how practicing a simple upperextremity motor task affects movement time for the task inpeople with PD.

2. Methods

2.1. Literature Search. The electronic databases used to findresearch that evaluated upper extremity motor learning inpeople with PD were CINAHL, EMBASE, PubMED, MED-LINE, PEDro, Proquest, PsycINFO, the Cochrane Databaseof Systematic Reviews, and Scopus. The comprehensivesearch used terms within the following categories: motorlearning, Parkinson’s disease, upper extremity, and time/speed/rate. The specific terms within categories are listed inTable 1.

The first four authors worked in pairs. Each pair wasrandomly assigned to search a set of databases and to selectarticles for screening. This initial search strategy resulted in127 articles.

2.2. Criteria for Inclusion in Systematic Review. Once the setof 127 articles was retrieved, the first four authors evaluatedthem. The title, abstract, and full content of all articleswere screened against the inclusion criteria, with each articleappraised by two of the first four authors. Based on the cri-teria, articles for inclusion in the meta-analysis were chosen.

Table 1: Search terms used for the meta-analysis.

Parkinson’sdisease

Upperextremity

Time/speed/rate

Practice Parkinson Arms Reaction time

Training PDUpperlimb

Serial reaction time

Sequential learning Parkinson’s Hand Reach time

Procedural learningParkinson

diseaseWrist Hand to mouth time

Motor skill learning Reaching Movement time

Skill learning Response time

Task performance Reaction speed

Task demand Serial reaction speed

Responseprogramming

Reach speed

Motor function Hand to mouth speed

Motor function loss Movement speed

Motor activity Response speed

Reaction rate

Serial reaction rate

Reach rate

Hand to mouth rate

Movement rate

Response rate

Where there was disagreement between members of the pairof reviewers, the fifth and sixth authors (S. J. Spaulding andM. E. Jenkins) were consulted, and a consensus was reached.Inclusion criteria were as follows: articles that were publishedbetween the beginning of included databases up to Septem-ber 2010, articles published in English, studies that examinedupper extremity motor learning in individuals with PD, stud-ies that included means and standard deviation or standarderror, studies that evaluated motor learning with time as anoutcome measure, and studies that had a control group.

Following the methodologies used by Siegert et al. [17],articles in the “grey literature,” such as conference proceed-ings or research published in Master’s or PhD theses, wereexcluded to avoid the use of evidence that had not been peerreviewed at the level of a journal article. After the applicationof the initial inclusion criteria, the authors had determinedthat 30 articles met all the criteria.

The authors then examined the experimental design ofthese 30 articles to determine research that provided pre/postmeasurements of movement time prior to and followingan intervention designed to elicit motor learning. Thefinal group of articles included five publications publishedbetween 1998 and 2009. Within those articles, there wereseven independent studies.

2.3. Data Extraction for Meta-Analysis. The first four authorsworking in pairs extracted the data from the seven indepen-dent studies. The following information was obtained forboth experimental and control groups in all studies: samplesize, pretraining mean, pretraining standard deviation orstandard error, posttraining mean, and posttraining standard


Table 2: Descriptive statistics of participants with PD in the included studies.

StudyAge mean

(SD)MMSE mean

(SD)Duration of PD inyears mean (SD)

Hoen and Yahrstage mean (SD)

UPDRS mean(SD)

Medication status(related to anti-Parkinsonian

medication)

Agostino et al. (2004) [21] 64.4 (6.3) >26 7.6 (3.1) N/A1 15.3 (4)(motor score)

On

Behrman et al. (2000) [22] 74 (7) 28 (1.6) 7 (4) 2.6 (0.5) N/A1 On

Majsak et al. (2008) [23] 70.4 (3.7) N/A 7.3 (7.9) 3 (0)33 (7.5)

(motor score)On

Marinelli et al. (2009)a [18] 60 (7.4) ≥27 8.4 (4.5) 2 to 2.5 N/A1 On

Marinelli et al. (2009)b [18] 57.9 (7.3) ≥27 2.1 (3.1) 1 to 2 N/A1 Off

Platz et al. (1998)a [4] 65.9 (8.3) 27.7 (1.6) 7.6 (2.4) 2.5 (0.5)8.0 (4)

Bradykinesiascore2

Off

Platz et al. (1998)b [4] 62.0 (14.6) 28.8 (1) 4.3 (1.8) 2.0 (.75)4.0 (3.5)

Bradykinesiascore

Off

1N/A indicates that the results were not available. SD: standard deviation.

2[24].Note: a and b are data from two different paradigms within one publication.c and d are data from two different experiments within one publication.

deviation. All time point values were documented immedi-ately following the intervention and late (in terms of timeafter practice) as defined by each individual study. Data wereextracted from text or figures, depending on how each articlepresented the data. If the resultant data were presented in afigure, each author, in the original pair of authors, extractedvalues, thus two measures were taken from the figure.The final value used was an average of the two authors’extracted numbers. Three studies reported both immediateand follow-up scores. When more than one follow-up periodwas measured, the authors chose to use the longest intervalbetween training and followup. For the purposes of thismeta-analysis, this period was termed late after training.Platz et al. [4] and Marinelli et al. [18] included two separatestudies in their articles. The studies had different numbers ofparticipants and different paradigms; thus, the results wereentered into the analysis separately.

2.4. Meta-Analysis. A meta-analysis was conducted usingthe program Comprehensive Meta-Analysis (CMA) [19].Hedge’s g, a measure of the standardized mean difference,was determined for the pre/postscores in each of the controlgroup and the group of individuals with PD. Hedge’s gaccounts for the overestimation of the population-standard-ized differences [20].

Because it could not be assumed that the people in thestudies were highly homogeneous in their characteristics, arandom effects model was used and provided a conservativeestimate of the differences between the groups in the individ-ual studies [20].

3. Results

A total of 58 individuals with PD and 56 participants with-out PD were included from the seven studies. Descriptive

statistics of all the subjects are included in Table 2. Descrip-tive statistics of the findings extracted from the studiesincluded in this meta-analysis are shown in Table 3. Table 4outlines the description of the motor learning paradigms inthe studies used in the meta-analysis.

Hedge’s g with a 95% confidence interval (CI) for each ofthe included studies is summarized in Table 5.

As seen in the forest plot representing the results for thecontrol group (Figure 1(a)), the point estimator of the overalleffect shows that participants without PD demonstrated im-provements in movement time. The point estimator of theoverall effect for individuals with PD did show improve-ments, but the changes were smaller and showed greater vari-ability than did the results of the control group (Figure 1(b)).The interval estimators of the overall effects (95% CI) foreach group overlapped. When comparing movement timesimmediately (early) posttraining to late posttraining, slowertimes of movement and larger 95% CI were evident for thelater posttraining time, for both groups.

4. Discussion

Although many studies have reported that motor learningoccurs in individuals with PD, not all studies have reportedimprovements [4]. Among studies that examine the acquisi-tion and retention of motor skills in PD, study sizes have beensmall, making conclusions less certain [6, 15]. In addition,tasks, duration of practice, and frequency of practice trialsare different between studies [6]. This meta-analysis wasable to overcome the heterogeneity issue by focusing onlyon studies of upper extremity movements and studies thatanalyzed improvements in movement time. Through theapplication of meta-analytic analysis, we were able to poolresults with heterogeneous methods and demonstrate a con-sistent reduction in movement time as a result of practice ofupper extremity reaching tasks.


Table 3: Descriptive statistics of studies of upper lime reach task.

Study

Control group Parkinson’s disease

Pre Immediate post Late post Pre Immediate post Late post

Mean time (SD) units: seconds Mean time (SD) units: seconds

Agostino et al. (2004) [21]N = 9 (PD)N = 7 (controls)

0.305 (0.026) 0.271 (0.035) 0.238 (0.246) 0.325 (0.286) 0.275 (0.750) 0.250 (0.394)

Behrman et al. (2000) [22]N = 15 (PD)N = 15 (controls)

0.183 (0.068) 0.106 (0.038) 0.111 (0.041) 0.200 (0.074) 0.130 (0.032) 0.134 (0.035)

Majsak et al. (2008) [23]N = 8 (PD)N = 8 (controls)

0.388 (0.062) 0.375 (0.058) 0.375 (0.035) 0.547 (0.110) 0.505 (0.095) 0.463 (0.047)

Marinelli et al. (2009)a [18]N = 5 (PD)N = 5 (controls)

0.440 (0.014) 0.430 (0.015) 0.440 (0.011) 0.430 (0.011)

Marinelli et al. (2009)b [18]N = 11(PD)N = 11 (controls)

0.425 (0.027) 0.415 (0.035) 0.400 (0.023) 0.415 (0.189)

Platz et al. (1998)c [4]N = 7 (PD)N = 7 (controls)

0.750 (0.138) 0.550 (0.072) 0.950 (0.051) 0.850 (0.080)

Platz et al. (1998)d [4]N = 8 (PD)N = 8 (controls)

0.750 (0.138) 0.620 (0.072) 0.950 (0.051) 0.865 (0.080)

N : number of subjects in each group, SD: standard deviation.Note: a and b data were extracted from two different paradigms within one publication. The first paradigm did not include cueing and the second did.c and d data were extracted from two different experiments within one publication.

Table 4: Description of the motor learning paradigms in the studies used in the meta-analysis.

Study Type of task Duration of practice Frequency of practice trials

Agostino et al. (2004) [21]Visually guided motor sequence infree space.

100 motor sequences trials.1 session/day (Monday to Friday).

2 weeks of 5 sessions/week.

Behrman et al. (2000) [22]Two simple sequential arm-reachingtasks between targets 12.7 cm apart.

120 reaction time trials. 1 session on each of 2 days.

Majsak et al. (2008) [23] Reaching a ball in front of person.5 blocks of 4 trials with blocksof stationary, moving, or dropball conditions.

90 minutes, approximately.1 session.

Marinelli et al. (2009)a [18]Reach on digitized tablet to arotating target from center.

48-second blocks of two tasks:with and without rotation.

1 session.

Marinelli et al. (2009)b [18]Reach on digitized tablet.Counterclockwise predicted.Clockwise not predicted.

90-second blocks of each oftwo tasks: predictable andunpredictable.

1 session.

Platz et al. (1998)c [4]Pointing from starting position totarget 20 cm away.

15 trials baseline, 100 trialspractice, and 15 trials witheach limb.

1 session.

Platz et al. (1998)d [4]Pointing from starting position totarget 20 cm away. Timing cuesprovided.

15 trials baseline, 100 trialspractice, and 15 trials witheach limb.

1 session.

Note: a and b are data from two different paradigms within one publication.c and d are data from two different experiments within one publication.

The results of the meta-analysis suggest that motor learn-ing in upper extremity function occurs in both neurologicallyhealthy controls and individuals with PD through practice ofupper extremity reaching tasks designed to reduce movementtime. This effect is present immediately after the training

period but also is sustained after a period of time althoughthe late effects are somewhat diminished. The control partic-ipants have a mild to moderate increased effect based on theirmean effect sizes compared to people with PD. However, thesubstantial overlap of confidence intervals would suggest that


Table 5: Effect sizes (as measured using Hedge’s g) with upper and lower 95% confidence intervals for the studies included in the meta-analysis and the resultant effect sizes. A negative value of the effect sizes is indicative of a reduction in the movement time.

(a)

AuthorsControl group Individuals with PD

Time of testing∗ Effect size (Hedge’s g) 95% CI Effect size (Hedge’s g) 95% CI

Agostino et al. (2004) [21] Immediate −0.937 −1.668 to −0.582 −0.177 −0.773 to 0.419

Behrman et al. (2000) [22] Immediate −1.233 −1.884 to −0.582 −1.031 −1.663 to −0.426

Majsak et al. (2008) [23] Immediate −0.192 −0.815 to 0.431 −0.361 −1.002 to 0.280

Marinelli et al. (2009)a [18]experiment 1

Immediate −0.551 −1.331 to 0.229 −0.727 −1.561 to 0.106

Marinelli et al. (2009)b [18]experiment 2

Immediate −0.265 −.955 to 0.425 −0.071 −0.746 to 0.604

Platz et al. (1998)c [4] study 1 Immediate −0.667 −1.197 to −0.156 −1.581 −2.400 to −0.863

Platz et al. (1998)d [4] study 2 Immediate −2.030 −2.873 to −1.186 −0.992 −1.571 to −0.414

Group immediate effect −0.814 −1.288 to −0.340 −0.698 −1.070 to −0.325

(b)

AuthorsControl group Individuals with PD

Time of testing∗ Effect size (Hedge’s g) 95% CI Effect size (Hedge’s g) 95% CI

Agostino et al. (2004) [21] Late −2.174 −3.339 to −1.009 −0.256 −0.857 to 0.346

Behrman et al. (2000) [22] Late −1.148 −1.778 to −0.517 −0.973 −1.565 to −0.381

Majsak et al. (2008) [23] Late −0.215 −0.839 to 0.410 −0.781 −1.506 to −0.056

Group late effect −1.028 −1.784 to 0.272 −0.665 −1.226 to −0.105

Overall effecte −0.875 −1.276 to −0.473 −0.688 −0.998 to −0.377

Note: a and b data were extracted from two different experiments within one publication. c and d data were extracted from two different training programswithin one publication. Effect size was corrected using Hedge’s g.eThe overall effect is the combination of the group immediate effect and the group late effect.∗Time of testing is indicated as either immediately following training (immediate) or following an interim period specified by each individual study (late).

Group by Study Time Hedges’s g and 95% CI

Immediate ImmediateImmediate Behram Immediate

Immediate ImmediateImmediate

Marinelli (a)Immediate

ImmediateMarinelli (b)

ImmediateImmediate ImmediateImmediate ImmediateAgostinoImmediateLate Behram LateLate Majsak Late

LateOverall

−4 −2 0 2 4

Majsak

Late Agostino Late

Results for control group

time point

Decrease time increase time

Platz (a)Platz (b)

(a)

Group by Study Time Hedges’s g and 95% CI

Immediate ImmediateImmediate Behram Immediate

Immediate ImmediateImmediate

Marinelli (a)Immediate

ImmediateMarinelli (b)

ImmediateImmediate ImmediateImmediate ImmediateAgostinoImmediateLate Behram LateLate Majsak Late

LateOverall

−4 −2 0 2 4

Majsak

Late Agostino Late

Results for individuals with PD

time point

Decreased time increased time

Platz (a)Platz (b)

(b)

Figure 1: Forest plots of all the included studies for the control group (a) and the individuals with PD (b) including the time the resultswere acquired, Hedge’s g, and 95% confidence interval (CI) for the control group. Each box and corresponding horizontal line represents theoverall mean and confidence intervals in the movement time. The area of each box is proportional to the inverse of that study’s variance. Thehorizontal line represents the 95% CI for each individual study. A diamond is used to depict overall mean effect size (center of the diamond)along with its CI (width of the diamond) [20].


both groups benefit from the practice in which they partici-pate.

Overall, these results are consistent with previous work insmall studies that demonstrate skill acquisition and retentionin people with PD in a variety of motor tasks. Such studieshave demonstrated acquisition and retention of motor skillsin varied upper extremity tasks not included in this meta-analysis such as serial reaction time tasks [25–27] and othersequential aiming movements [7, 9, 13]. Furthermore, motorlearning studies in people with PD have demonstrated im-provement in balance and lower extremity function throughpractice [10, 28–30].

In addition, motor learning effect, demonstrated by im-provement in movement time, was smaller among individ-uals with PD. This is not particularly surprising, given therole of the basal ganglia in both acquisition of motor taskskill and in consolidation of automatic movements [2, 3, 31].As evidence of the potential alterations of brain activityin persons with PD during task learning, functional MRIstudies in individuals with PD have demonstrated thatgreater areas of the brain are activated during initial learningof a task and particularly during the repetition of a learnedmovement in PD compared to healthy controls [31].

4.1. Rehabilitation Implications. A number of differenceswere identified in the experimental methodologies of thestudies from which data were extracted to conduct this meta-analysis. There was variability among the duration and fre-quency of practice as well as the types of tasks. These dif-ferences preclude the authors from determining that there isone type of practice that was more effective to improve upperextremity performance. However, one can conclude thatpractice in general is beneficial and the manipulation of prac-tice parameters is worthy of further study. Interestingly, evenin the studies in which the individuals were off dopaminereplacement medication [4, 18], there was a decrease inmovement time, suggesting that there could potentially be arehabilitation program that would benefit people with PD,even if medication effectiveness was suboptimal for somereason. Yet, current studies suggest that dopamine replace-ment medication may have a deleterious effect on motorlearning [32].

4.2. Limitations of the Study. A limitation of the presentmeta-analysis is the small number of studies that the authorswere able to include, but to the best of our knowledge, allof the available studies of simple reaching tasks reportingmovement time as an outcome were incorporated. There aremore studies evaluating practice, but they were heteroge-neous in their tasks or in their outcome measures; therefore,they did not meet our inclusion criteria, and the datacould not be included in this meta-analysis. Additionally, thesample sizes of the included studies were small, affecting thegeneralizability of this meta-analysis [33].

4.3. Recommendations for Future Research. Current literaturein this area typically examines one single task or movement.Future research might best examine the generalizability of

the effects of practice to other tasks and areas of rehabilita-tion. Conclusions from a broader range of tasks could leadto the use of programs that are directly related to movementsneeded for daily functioning. Finally, future motor skill ac-quisition research should further examine the effects of var-ied practice parameters in more diverse samples of personswith PD.

5. Conclusions

Results from this pooling of data from various studies pro-vide evidence that upper extremity movement time canbe improved through the use of practice of reaching tasksin persons with PD, albeit potentially to a lesser extentthan is shown in individuals with no neurological prob-lems. The collective interpretation of this meta-analysisindicates that practice of relevant motor tasks targeted atmaximizing acquisition and retention improved movementspeed.

References

[1] R. Schmidt and T. Lee, Motor Control and Learning: A Behav-ioral Emphasis, Human Kinetics Books, Champaign, Ill, USA,1999.

[2] J. Doyon, “Motor sequence learning and movement disor-ders,” Current Opinion in Neurology, vol. 21, no. 4, pp. 478–485, 2008.

[3] V. Puttemans, N. Wenderoth, and S. P. Swinnen, “Changes inbrain activation during the acquisition of a multifrequencybimanual coordination task: from the cognitive stage toadvanced levels of automaticity,” Journal of Neuroscience, vol.25, no. 17, pp. 4270–4278, 2005.

[4] T. Platz, R. G. Brown, and C. D. Marsden, “Training improvesthe speed of aimed movements in Parkinson’s disease,” Brain,vol. 121, no. 3, pp. 505–514, 1998.

[5] M. Hallett and S. Khoshbin, “A physiological mechanism ofbradykinesia,” Brain, vol. 103, no. 2, pp. 301–314, 1980.

[6] A. Nieuwboer, L. Rochester, L. Muncks, and S. P. Swinnen,“Motor learning in Parkinson’s disease: lmitations and poten-tial for rehabilitation,” Parkinsonism and Related Disorders, vol.15, supplement 3, pp. S53–S58, 2009.

[7] H. Mochizuki-Kawai, S. Mochizuki, and M. Kawamura, “Aflexible sequential learning deficit in patients with Parkinson’sdisease: a 2 × 8 button-press task,” Experimental BrainResearch, vol. 202, no. 1, pp. 147–153, 2010.

[8] L. Rochester, K. Baker, V. Hetherington et al., “Evidence formotor learning in Parkinson’s disease: acquisition, automatic-ity and retention of cued gait performance after training withexternal rhythmical cues,” Brain Research, vol. 1319, pp. 103–111, 2010.

[9] A. L. Smiley-Oyen, K. A. Lowry, and J. P. Kerr, “Planning andcontrol of sequential rapid aiming in adults with Parkinson’sdisease,” Journal of Motor Behavior, vol. 39, no. 2, pp. 103–114,2007.

[10] R. T. Jessop, C. Horowicz, and L. E. Dibble, “Motor learningand Parkinson disease: refinement of movement velocity andendpoint excursion in a limits of stability balance task,”Neurorehabilitation and Neural Repair, vol. 20, no. 4, pp. 459–467, 2006.


[11] W. D. Byblow, J. J. Summers, G. N. Lewis, and J. Thomas, “Bi-manual coordination in Parkinson’s disease: deficits in move-ment frequency, amplitude, and pattern switching,” MovementDisorders, vol. 17, no. 1, pp. 20–29, 2002.

[12] R. Inzelberg, M. Plotnik, T. Flash, E. Schechtman, I. Shahar,and A. D. Korczyn, “Mental and motor switching in Parkin-son’s disease,” Journal of Motor Behavior, vol. 33, no. 4, pp.377–385, 2001.

[13] M. A. Stephan, B. Meier, S. W. Zaugg, and A. Kaelin-Lang,“Motor sequence learning performance in Parkinson’s diseasepatients depends on the stage of disease,” Brain and Cognition,vol. 75, no. 2, pp. 135–140, 2011.

[14] C. H. Lin, K. J. Sullivan, A. D. Wu, S. Kantak, and C. J. Win-stein, “Effect of task practice order on motor skill learning inadults with Parkinson disease: a pilot study,” Physical Therapy,vol. 87, no. 9, pp. 1120–1131, 2007.

[15] S. P. Swinnen, M. Steyvers, L. van den Bergh, and G. E. Stel-mach, “Motor learning and Parkinson’s disease: refinementof within-limb and between-limb coordination as a result ofpractice,” Behavioural Brain Research, vol. 111, no. 1-2, pp. 45–59, 2000.

[16] J. G. Nutt, E. S. Lea, L. van Houten, R. A. Schuff, and G. J.Sexton, “Determinants of tapping speed in normal controlsubjects and subjects with Parkinson’s disease: differing effectsof brief and continued practice,” Movement Disorders, vol. 15,no. 5, pp. 843–849, 2000.

[17] R. J. Siegert, M. Weatherall, K. D. Taylor, and D. A. Abernethy,“A meta-analysis of performance on simple span and morecomplex working memory tasks in Parkinson’s disease,”Neuropsychology, vol. 22, no. 4, pp. 450–461, 2008.

[18] L. Marinelli, D. Crupi, A. Di Rocco et al., “Learning and con-solidation of visuo-motor adaptation in Parkinson’s disease,”Parkinsonism & Related Disorders, vol. 15, no. 1, pp. 6–11,2009.

[19] M. Borenstein, L. V. Hedges, J. P. T. Higgins, and H. R. Roth-stein, Comprehensive Meta-Analysis Program, Biostat, Engle-wood Cliffs, NJ, USA, 2005.

[20] M. Borenstein, L. V. Hedges, J. P. T. Higgins, and H. R. Roth-stein, Introduction to Meta-Analysis, John Wiley & Sons,Chichester, UK, 2009.

[21] R. Agostino, A. Curra, G. Soldati et al., “Prolonged practice isof scarce benefit in improving motor performance in Parkin-son’s disease,” Movement Disorders, vol. 19, no. 11, pp. 1285–1293, 2004.

[22] A. L. Behrman, J. H. Cauraugh, and K. E. Light, “Practice asan intervention to improve speeded motor performance andmotor learning in Parkinson’s disease,” Journal of the Neuro-logical Sciences, vol. 174, no. 2, pp. 127–136, 2000.

[23] M. J. Majsak, T. Kaminski, A. M. Gentile, and A. M. Gordon,“Effects of a moving target versus a temporal constraint onreach and grasp in patients with Parkinson’s disease,” Experi-mental Neurology, vol. 210, no. 2, pp. 479–488, 2008.

[24] S. Fahn and R. Elton, “Updrs development committee. Unifiedparkinson’s disease rating scale,” in Recent Developments inParkinson’s Disease, S. Fahn, C. Marsden, D. Calne, and M.Goldste, Eds., pp. 153–163, Macmillan, Florham Park, NJ,USA, 1987.

[25] I. van Tilborg and W. Hulstijn, “Implicit motor learning inpatients with Parkinson’s and Alzheimer’s disease: differencesin learning abilities?” Motor Control, vol. 14, no. 3, pp. 344–361, 2010.

[26] E. D. Stefanova, V. S. Kostic, L. Ziropadja, M. Markovic, andG. G. Ocic, “Visuomotor skill learning on serial reaction time

task in patients with early Parkinson’s disease,” MovementDisorders, vol. 15, no. 6, pp. 1095–1103, 2000.

[27] L. Wilkinson and M. Jahanshahi, “The striatum and proba-bilistic implicit sequence learning,” Brain Research, vol. 1137,no. 1, pp. 117–130, 2007.

[28] M. K. Mak, E. C. Wong, and C. W. Hui-Chan, “Quantitativemeasurement of trunk rigidity in parkinsonian patients,”Journal of Neurology, vol. 254, no. 2, pp. 202–209, 2007.

[29] J. Michel, D. Benninger, V. Dietz, and H. J. van Hedel, “Obsta-cle stepping in patients with Parkinson’s disease. Complexitydoes influence performance,” Journal of Neurology, vol. 256,no. 3, pp. 457–463, 2009.

[30] M. Jobges, G. Heuschkel, C. Pretzel, C. Illhardt, C. Renner, andH. Hummelsheim, “Repetitive training of compensatory steps:a therapeutic approach for postural instability in Parkinson’sdisease,” Journal of Neurology, Neurosurgery and Psychiatry,vol. 75, no. 12, pp. 1682–1687, 2004.

[31] M. Carbon and D. Eidelberg, “Functional imaging of sequencelearning in Parkinson’s disease,” Journal of the NeurologicalSciences, vol. 248, no. 1-2, pp. 72–77, 2006.

[32] Y. Kwak, M. L. T. M. Muller, N. I. Bohnen, P. Dayalu, and R.D. Seidler, “Effect of dopaminergic medications on the timecourse of explicit motor sequence learning in parkinson’sdisease,” Journal of Neurophysiology, vol. 103, no. 2, pp. 942–949, 2010.

[33] E. DePoy and L. N. Gitlin, Introduction to Research: Under-standing and Applying Multiple Strategies, Elsevier Mosby,Philadelphia, Pa, USA, 2005.


Research Article

Lack of Short-Term Effectiveness ofRotating Treadmill Training on Turning in People withMild-to-Moderate Parkinson’s Disease and Healthy Older Adults:A Randomized, Controlled Study

Marie E. McNeely1, 2 and Gammon M. Earhart1, 3, 4

1 Program in Physical Therapy, Washington University in St. Louis, St. Louis, MO 63108, USA2 Program in Neuroscience, Division of Biology and Biomedical Sciences, Washington University in St. Louis, St. Louis, MO 63110,USA

3 Department of Anatomy and Neurobiology, Washington University in St. Louis, St. Louis, MO 63110, USA4 Department of Neurology, Washington University in St. Louis, St. Louis, MO 63110, USA


Received 27 July 2011; Accepted 5 September 2011


Copyright © 2012 M. E. McNeely and G. M. Earhart. This is an open access article distributed under the Creative CommonsAttribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work isproperly cited.

Since turning is often impaired in Parkinson’s disease (PD) and may lead to falls, it is important to develop targeted treatmentstrategies for turning. We determined the effects of rotating treadmill training on turning in individuals with PD. This randomizedcontrolled study evaluated 180◦ in-place turns, functional turning (timed-up-and-go), and gait velocity before and after 15 minutesof rotating treadmill training or stepping in place in 26 people with PD and 27 age-matched controls. A subset of participants withPD (n = 3) completed five consecutive days of rotating treadmill training. Fast as possible gait velocity, timed-up-and-go time,180◦ turn duration, and steps to turn 180◦ were impaired in PD compared to controls (P < 0.05) and did not improve followingeither intervention (P > 0.05). Preferred pace gait velocity and timing of yaw rotation onset of body segments (head, trunk, pelvis)during 180◦ turns were not different in PD (P > 0.05) and did not change following either intervention. No improvements in gaitor turning occurred after five days of rotating treadmill training, compared to one day. The rotating treadmill is not recommendedfor short-term rehabilitation of impaired in-place turning in the general PD population.

1. Introduction

Parkinson’s disease (PD) is a progressive neurodegenerativedisease resulting in a variety of motor symptoms. Individualswith PD frequently experience difficulty with gait and turn-ing, with more than half reporting difficulty turning [1–3]which may result in falls and serious injuries [4]. Symptomsof PD are treated using various therapeutic approaches;however, there are currently no effective treatment optionsthat specifically target turning difficulty. Turning difficulties,including increased time to turn and increased number ofsteps to turn, are present even when individuals with PD areon PD medications [5–10].

Stepping in place on the rotating treadmill has been rec-ommended as a possible rehabilitation option for those with

PD [11]. After stepping in place on the rotating treadmill,healthy controls and people with PD show a rotationaladaptation response known as podokinetic after-rotation[12–14]. The kinematics of podokinetic after-rotation aresimilar to those seen during normal in-place turning [11]. Ithas been suggested that the rotating treadmill may improveturns by serving as an external cue to promote the correctmotor programs for successful turning [11].

Immediately after stepping in place on a rotating diskfor a total of 15 minutes on one day, turning performancewas improved in two people with PD on medicationwho also experienced freezing of gait during turns [15].Specifically, there were fewer freezing events, reduced timeto turn, less variable vastus lateralis muscle activity, andreduced coactivation of bilateral vastus lateralis muscles


Table 1: Participant demographics.

CN CN PD PD PD

1-day train 1-day step 1-day train 1-day step 5-day train

Total n 14 13 12 14 3

Age (yrs) 65.3 ± 11.3 70.1 ± 11.4 69.1 ± 9.7 70.0 ± 11.2 69.0 ± 17.0

Males/females 9/5 5/8 8/4 8/6 2/1

Disease duration (yrs) NA NA 8.5 ± 4.9 6.6 ± 5.5 8.7 ± 5.7

UPDRS-III NA NA 25.8 ± 10.3 26.9 ± 7.4 27.7 ± 20.6

H & Y stage NA NA 2.1 ± 0.4 2.1 ± 0.7 2.0 ± 0.9

Values are means ± SDs.

[15]. It remains unclear whether improvements seen withrotating treadmill training are specific to people with PDwho experience freezing, nor do we know if improvementsoccur in other aspects of turning. If turning improvementsoccurred in a more diverse group of individuals with PD,the rotating treadmill would potentially be relevant as arehabilitation tool. Our aim was to conduct a randomizedcontrolled study examining the effects of rotating treadmilltraining on in-place turning in a larger, more representativegroup of individuals with PD and age-matched healthycontrols. This study also includes control exercise groupsfor PD and healthy older adult participants. These controlgroups stepped in place on the floor for an amount oftime equal to the treadmill training performed by the othergroups. We hypothesized that turning would improve inindividuals with PD following rotating treadmill training,while turning would likely remain unchanged for all healthyolder adults and for those with PD who stepped in place onthe floor.

2. Methods

2.1. Participants. We recruited 29 participants from theMovement Disorders Center at Washington UniversitySchool of Medicine who had been diagnosed with idiopathicParkinson’s disease according to standard criteria [16]. Wealso recruited 28 older adults without PD. People with PDwere recruited if they were taking medication for PD, wereambulatory, did not have deep brain stimulators implanted,had no history or symptoms of other neurological diseases,and had no recent surgeries or injuries affecting walkingor turning. Those with PD were tested approximately onehour after their last dose of PD medication. Of the 29people with PD recruited, 3 did not complete the studydue to fatigue. These individuals were excluded from allsubsequent analyses. Older adults without PD were recruitedif they were ambulatory, had no history or symptomsof neurological diseases, and had no recent surgeries orinjuries affecting walking or turning. Of the 28 controlsrecruited, one did not complete the study due to fatigue andwas excluded. Demographics for included participants areshown in Table 1. All participants provided written informedconsent prior to participation, and this study was approvedby the Washington University School of Medicine HumanResearch Protection Office.

2.2. Experimental Design. All participants with and withoutPD completed testing on one day, and a small subset ofparticipants with PD (n = 3) returned for an additionalfive consecutive days of training and testing. Training andtesting sessions were identical for the one-day and five-day portions of the study. Surgical skin pens were usedto ensure consistent placement of reflective markers acrossdays. The Movement Disorders Society Unified Parkinson’sDisease Rating Scale motor subscale (MDS-UPDRS-III) wasgiven to all participants prior to testing to assess movementimpairments [17].

2.3. Intervention. Participants with and without PD wererandomly assigned (computer-based algorithm) to an inter-vention condition (rotating treadmill (Train) or stepping inplace (Step)). Those in the Train condition were asked tostep in place on the perimeter of a rotating disk built intothe floor (120 cm diameter, Neuro Kinetics, Inc., Pittsburgh,Pa) as it rotated approximately 45◦/sec either clockwise orcounterclockwise. The direction of treadmill rotation wasselected for each participant to train turns in the worsedirection (i.e., the direction requiring greater time to turnin place 180◦), where clockwise rotation trained left turnsand counterclockwise rotation trained right turns. For thefive-day training sessions, right turns were trained for allparticipants. Participants walked on the rotating treadmillfor a total of 15 minutes, divided into 5-minute blocks withinterspersed 5-minute rest periods [14]. Those in the Stepcondition experienced a similar amount of physical activityby stepping in place at a self-selected pace on the stationaryground for a total of 15 minutes, divided into 5-minuteblocks with interspersed 5-minute rest periods.

2.4. Data Collection and Analysis. Turning and walking wereassessed in two separate blocks, once before (PRE) and onceafter (POST) the assigned intervention. In each testing block,we examined gait, functional turning while walking, and in-place turns of 180◦. Gait was assessed using a 4.8 m GAITRiteinstrumented walkway (CIR Systems, Havertown, Pa) todetermine if the interventions had any effects on gait. TheGAITRite calculated gait velocity in six walking trials in eachblock: three at a participant’s preferred pace and three as fastas possible. Functional turning ability was assessed using thetimed-up-and-go (TUG) test where participants rise from achair, walk three meters, turn 180◦, walk three meters back


0

25

50

75

100

125

150

175

200

225

Train Step Train Step

PD CN

Vel

ocit

y(c

m/s

)

Preferred pace gait

(a)


PD CN

0

25

50

75

100

125

150

175

200

225

Vel

ocit

y(c

m/s

)

Fast pace gait

P = 0.034

PREPOST

(b)

Figure 1: Gait Velocity. Mean preferred pace gait velocity (a) and fast as possible gait velocity (b) for PD and controls in the rotating treadmilltraining group (Train) and the stepping-in-place group (Step). Brackets indicate a significant group effect. Error bars are SDs.

to the chair, and sit down. Participants completed the TUGsix times in each block, with instruction to turn left or righton each trial (3 trials of each, randomized). The time tocomplete this task was measured using a stopwatch.

To assess in-place turning, we instructed participantsto stand in the middle of the room and turn 180◦ to theright or left, 10 times in each direction (randomized), toface the wall behind them. Kinematic data were recordedat 100 Hz during these turns using an eight-camera 3-Dmotion capture system (Motion Analysis Corp., Santa Rosa,Calif). Thirty-four reflective markers were placed on eachparticipant: four on the head (forehead, back of the head,and above the left and right ears), seven on the trunk (leftand right acromion processes, right scapula, seventh cervicalvertebra, tenth thoracic vertebra, sternal notch, and xiphoidprocess), five on the pelvis (left and right anterior superioriliac spines, left and right posterior superior iliac spines, andsacrum), and nine on each leg (greater trochanter, anteriorthigh, lateral femoral condyle, tibial tuberosity, middle tibia,lateral malleolus, calcaneus, navicular, base of the secondmetatarsal).

The Motion Monitor software (Innovative Sports Train-ing, Inc., Chicago, Ill) was used to create body segmentmodels for the head, trunk, pelvis, and feet based onmarker positions and to export segment rotation data.Custom written Matlab programs were used (MathWorks,Inc., Natick, Mass) to determine rotation onsets and offsetsfor the head, trunk, pelvis, and feet body segments, usinga 5◦ yaw plane rotation threshold criterion. The onset ofrotation of the foot used for the first step was designated asthe turn onset, and the offset of rotation of the foot used forthe last step was designated as the turn offset. To quantifybody segment rotation sequences during turn initiation, we

examined the timing of the onsets of the head (HTO), trunk(TTO), and pelvis (PTO) relative to turn onset. All onsettimes were expressed as a percentage of the first gait cycleof the turn. To assess overall turn performance, we alsodetermined the number of steps used to complete each turnand turn duration. Kinematic, TUG, and gait velocity datawere averaged across trials within each participant for thePRE block and for the POST block.

2.5. Statistical Analyses. Our primary variables of interestwere functional turning ability (TUG), 180◦ in-place turnduration, and normalized rotation onset of the head, trunk,and pelvis relative to turn onset, to quantify timing of bodysegment rotations during turn initiation. Secondarily, welooked at number of steps to turn 180◦. We also examinedvelocity during preferred pace and fast as possible gait todetermine if the interventions, specifically the rotating tread-mill, impacted gait. In the Train groups, we trained the worseturn direction, and the other turn direction was untrained.In order to similarly compare turn performance for theStep groups, we designated the worse turn direction and thebetter turn direction based on turn durations. For statisticalcomparisons across groups, the trained direction of the Traingroup was compared with the worse direction of the Stepgroup, and the untrained direction of the Train group wascompared with the better direction of the Step group. Wewere primarily interested in the trained direction, as turns inthis direction were expected to be affected by rotating tread-mill training. Separate RM-ANOVAs were run (RM factor:Time; between subjects factors: Condition, Group) for ourprimary variables of interest for the trained/worse directions.We also ran RM-ANOVAs for step number and gait velocity.Only 3 participants with PD completed the 5-day training


0

4

8

12

16

20


PD CN

Tim

e(s

)

Timed-up-and-go (TUG)

P = 0.026

(a)


PD CN

Tim

e(s

)

0

1

2

3

4

5

6

7

8

180◦ turn duration

P < 0.001

(b)


PD CN

0

2

4

6

8

10

12

14

Nu

mbe

rof

step

s

Steps to turn 180◦

P = 0.001

PREPOST

(c)

Figure 2: Functional Turning and Turn Performance. Mean timed-up-and-go time (a), 180◦ in-place turn duration (b), and number ofsteps to turn 180◦ in-place (c) for PD and controls in the rotating treadmill training group (Train) and the stepping in place group (Step).Brackets indicate significant group effects. Error bars are SDs.

component, so formal statistical tests were not used. How-ever, we examined POST data from Day 1 and Day 5 forprimary variables of interest, step number, and gait velocityto determine trends. Secondarily, we also evaluated variablessimilarly for the untrained/better directions.

3. Results

3.1. 1-Day Training. The PD and CN groups did not differsignificantly in age (P = 0.503). The PD Train and Stepgroups had similar ages, MDS-UPDRS-III scores, and diseasedurations (P > 0.05). Similar ages and MDS-UPDRS-IIIscores were also seen across CN Train and Step groups (P >0.05). There were no significant differences in any variablesat baseline for turns in the trained/worse direction or turnsin the untrained/better direction (P > 0.05). There werealso no significant differences between those with left asthe trained/worse turn direction and those with right as thetrained/worse direction (P > 0.05), so data were combinedfor analysis.

3.1.1. Gait and Functional Turning. GAITRite data for twoparticipants (1PD, 1CN) were lost due to hard drive failure,but all remaining data for these participants was includedin analyses. The mean velocity data are shown for PDand controls before and after the assigned intervention inFigure 1. There were no significant effects of Condition(f(1,47) = 2.57, P = 0.12) or Group (f(1,47) = 1.40, P =0.24), nor any significant interaction effects (P > 0.05)for preferred pace gait velocity (Figure 1(a)). There was atrend towards an effect of Time (f(1,47) = 3.79, P = 0.06),with individuals tending to demonstrate higher preferredpace gait velocity POST intervention. For fast as possiblegait velocity (Figure 1(b)), there were no significant effectsof Time (f(1,47) = 0.25, P = 0.62) or Condition (f(1,47)= 3.28, P = 0.08), nor any significant interaction effects(P > 0.05). There was a significant effect of Group (f(1,47)= 4.77, P = 0.034), with PD walking slower than CN.

For TUG where the turn component was in thetrained/worse direction (Figure 2(a)), there were no signif-icant effects of Time (f(1,49) = 0.31, P = 0.58) or Condition(f(1,49) = 3.43, P = 0.070), nor any significant interactions(P > 0.05). There was a significant Group effect (f(1,49) =


020406080

100120140160180200

Yaw

rota

tion

ampl

itu

de(d

eg)

020406080

100120140160180200

Yaw

rota

tion

ampl

itu

de(d

eg)

PR

EP

OST

0 1 2 3 4 5 6

Time (s)

0 1 2 3 4 5 6

Time (s)

0 1 2 3 4 5 6

Time (s)

0 1 2 3 4 5 6

Time (s)

PD-Train PD-Step CN-Train CN-Step

HeadTrunk

PelvisRFoot

LFoot

Figure 3: Turn Kinematics Examples. Representative traces of yaw plane rotations of individual body segments during a single 180◦ in-placeturn in one PD-Train, one PD-Step, one CN-Train, and one CN-Step participant before (PRE) and after (POST) intervention.

5.25, P = 0.026), with PD requiring more time to completethe TUG, turning to the trained/worse direction, comparedto controls. Results were similar for the untrained/betterdirection.

3.1.2. Turn Kinematics. For 180◦ turn duration in thetrained/worse direction (Figure 2(b)), there were no signifi-cant effects of Time (f(1,49) = 0.025, P = 0.62) or Condition(f(1,49) = 0.99, P = 0.33), nor any significant interactions(P > 0.05). There was a significant Group effect (f(1,49) =15.95, P < 0.001), with PD turning slower than CN. For theuntrained/better direction, results were similar.

For steps to turn 180◦ in the trained/worse direction(Figure 2(c)), there were no significant effects of Time(f(1,49) = 2.15, P = 0.15) or Condition (f(1,49) = 1.33,P = 0.25), nor any significant interactions (P > 0.05). Therewas a significant Group effect (f(1,49) = 13.71, P = 0.001),with PD requiring more steps to turn. Similar results wereseen for the untrained/better direction.

For body segment (head, trunk, pelvis) rotation onsetsrelative to turn onset for turns in the trained/worse direction,there were no significant effects of Time (f(3,47) = 1.26,P = 0.30), Condition (f(3,47) = 0.60, P = 0.62), or Group(f(3,47) = 1.48, P = 0.23), nor any significant interactions(P > 0.05). Figure 3 shows representative sample tracesfrom single individuals in the PD-Train (a), CN-Train (b),PD-Step (c), and CN-Step (d) groups. Mean body segmentrotation onsets are shown in Figure 4 for the PD and CNgroups for the Train (a, b) and Step (c, d) conditionsbefore and after intervention. In all groups, the sequence ofrotation onsets of body segments was head first, followedby trunk, and then pelvis. Comparisons for body segment

rotation onsets relative to turn onset were similar for theuntrained/better direction.

3.2. 5-Day Training. A small subset of the original groupof participants returned for 5 consecutive days of rotatingtreadmill training. All three individuals who returned for fiveconsecutive days of training were able to tolerate the trainingprogram. On average, gait velocity, TUG, turn duration,steps to turn, and body segment rotation onsets relative toturn onset were very similar following a single session oftraining (Day 1 POST), compared to after five sessions oftraining (Day 5 POST) in either the trained or untraineddirection. Table 2 shows baseline data from Day 1 prior totraining, as well as from Day 1 and Day 5 after training forthe trained direction.

4. Discussion

Difficulty with turning is common in individuals with PD,and the development of therapeutic approaches that targetturn deficits might reduce the occurrence of falls and seriousinjuries in these individuals. As a result, it is important toevaluate potential treatment strategies in individuals with PDwho demonstrate a range of turning ability and might benefitfrom these treatment options.

Contrary to our initial hypotheses, we did not seeimprovements in turning in those with PD following oneday or five consecutive days of rotating treadmill training.For most turning variables, group effects indicated turningwas impaired in those with PD ON medication, compared tocontrols, as has been previously reported [5–10].


PD-Train

−50 −40 −30 −20 −10 0

PTO

TTO

HTO

(a)

−50 −40 −30 −20 −10 0

PTO

TTO

HTO

CN-Train

(b)

−50 −40 −30 −20 −10 0

PTO

TTO

HTO

First stride time (%)

PD-Step

(c)

PTO

TTO

HTO

−50 −40 −30 −20 −10 0

First stride time (%)

CN-Step

PREPOST

(d)

Figure 4: Onsets of Body Segment Rotations. Mean yaw rotation onset times of the head (HTO), trunk (TTO), and pelvis (PTO) relative tothe turn onset (i.e., first foot rotation) are expressed as a percentage of the first stride of the turn for PD-Train (a), CN-Train (b), PD-Step(c), and CN-Step (d). Error bars are SDs.

Table 2: Five-day rotating treadmill training results for the trained direction.

Day 1 PRE Day 1 POST Day 5 POST

Fwd gait velocity (cm/sec) 105.1 ± 24.1 109.8 ± 18.3 110.4 ± 24.6

Fast gait velocity (cm/sec) 149.3 ± 25.5 149.5 ± 28.0 152.5 ± 26.2

TUG (sec) 12.0 ± 3.3 12.0 ± 3.5 12.1 ± 2.7

Turn duration (sec) 2.7 ± 0.8 2.6 ± 0.9 2.5 ± 0.8

Steps to turn 4.6 ± 0.5 4.6 ± 0.6 4.4 ± 0.5

NHTO (% Gait Cycle) −30.7 ± 11.4 −24.3 ± 7.6 −23.5 ± 12.9

NTTO (% Gait Cycle) −25.1 ± 9.5 −24.7 ± 8.4 −22.5 ± 13.7

NPTO (% Gait Cycle) −24.0 ± 10.7 −22.8 ± 8.8 −20.1 ± 13.1

Values are means ± SDs.

Interestingly, the body segment rotation onset patternswe observed were similar between those with PD andcontrols. All groups initiated turns with the head, followedby the trunk, pelvis, and foot. Controls display this top-downrotation sequence during turns while walking [8, 18–24],

as well as in-place turns [25]. In contrast, those with PDhave been reported to display more simultaneous rotationof the head, trunk, and pelvis during turning while walking,including a pronounced delay in initiation of head rotation


[5, 8, 10]. We did not see this kinematic pattern during in-place turns in people with PD. A previous report of impairedbody segment rotation patterns in PD during in-place turnsindicated there were no significant differences in rotationonset time between body segments [25], though averageonset times of each segment followed a top-down sequence,similar to results of the present study.

Turning and gait velocity did not systematically worsen inany intervention group from PRE to POST. Also, participantswere permitted to rest as often as necessary during testingand had mandatory five-minute rests between the five-minute blocks of treadmill training. As a result, we thinkour results were likely not confounded by fatigue. Further,we likely tested a higher-functioning group of people withPD than some prior studies, since preferred pace gaitvelocity did not differ between PD and controls. Despitesmaller differences in gait performance, we detected distinctimpairments in turn performance compared to controls,confirming turning impairments can be present in thosewith relatively normal gait [5]. This highlights the need forrehabilitation to address turning deficits even in the earlystages of disease progression.

The most notable difference between the present studyand the previous rotating treadmill study [15] is the focus ofthe earlier study on individuals with severe freezing. Peoplewith PD with freezing of gait can improve motor perfor-mance via external cues, and cued exercise interventions havebeen used to improve locomotion in PD with freezing of gait.In one study, robot-assisted gait training improved freezingof gait frequency, as well as gait velocity, stride length,coordination, and rhythmicity in those with PD with freezingof gait [26]. It is possible that for individuals with moresevere PD or with severe freezing, training on the rotatingtreadmill may help make the appropriate turning motorpatterns more automatic. This might in turn facilitate theirimpaired task switching [27–29], reducing the frequencyand severity freezing during turns. The present study onlyincluded 9 individuals with freezing of gait, as defined byreports of freezing at least once per week on item threeof the Freezing of Gait Questionnaire [30]; however, onlyone individual with freezing was randomly assigned to therotating treadmill training group, so comparisons betweenthose with and without freezing could not be made. Thesmall overall sample size and the fact that only one personwith freezing trained on the rotating treadmill are limitationsof the study and warrant careful interpretation of the data, asthe study may have been underpowered to detect interactioneffects.

Another limitation of the study is that the sessions wereof relatively low intensity and were few in number. It maybe that more intense rotating treadmill training sessionsor increased number of sessions may result in detectablechanges, as previous traditional treadmill studies reportimprovements after completion of 10–28 training sessions of20–30 minutes each [26, 31]. However, there are also reportsof acute effects on gait from just one session of traditionaltreadmill training [32–34]. Another possibility is that therotating treadmill may be more useful for people with PDwhen combined with other cueing strategies. Combining

traditional treadmill training with auditory and visual cuesimproved gait speed, maximum distance walked in sixminutes, and score on the Freezing of Gait Questionnaire inone study of people with PD [31].

5. Conclusions

Fifteen minutes of rotating treadmill training alone onone day or for five consecutive days did not affect turnperformance in PD. As a result, this type of training isunlikely to serve as an effective short-term rehabilitationstrategy for many individuals with PD. However, futurestudies should determine whether rotating treadmill trainingmay improve turning impairments with longer trainingparadigms or when combined with other external cues, aswell as assess its effects on performance of turns whilewalking in addition to the in-place turning studied here.

Acknowledgments

Research support was provided by the National Institute ofHealth/National Institute of Neurological Disease and StrokeAward number F31 NS071639, the National Institute ofHealth/National Center for Medical Rehabilitation ResearchAward number R01 HD056015, the American Parkinson’sDisease Association (APDA) Advanced Center for PDResearch at Washington University School of Medicine, andthe Greater Saint Louis Chapter of the APDA. The authorswould like to thank Samantha Herriott and Vanessa Heil-Chapdelaine for assisting with data processing, as well asRyan Duncan, Corey Lohnes, Daniel Peterson, and JohnMichael Rotello for assisting during data collection.

References


[2] A. Nieuwboer, W. de Weerdt, R. Dom, and E. Lesaffre,“A frequency and correlation analysis of motor deficits inParkinson patients,” Disability and Rehabilitation, vol. 20, no.4, pp. 142–150, 1998.

[3] E. L. Stack, A. M. Ashburn, and K. E. Jupp, “Strategies used bypeople with Parkinson’s disease who report difficulty turning,”Parkinsonism and Related Disorders, vol. 12, no. 2, pp. 87–92,2006.



[6] M. E. Morris, F. Huxham, J. McGinley, K. Dodd, and R. Iansek,“The biomechanics and motor control of gait in Parkinsondisease,” Clinical Biomechanics, vol. 16, no. 6, pp. 459–470,2001.


[7] E. Stack and A. Ashburn, “Dysfunctional turning in Parkin-son’s disease,” Disability and Rehabilitation, vol. 30, no. 16, pp.1222–1229, 2008.

[8] I. Carpinella, P. Crenna, E. Calabrese et al., “Locomotorfunction in the early stage of Parkinson’s disease,” IEEETransactions on Neural Systems and Rehabilitation Engineering,vol. 15, no. 4, pp. 543–551, 2007.

[9] J. E. Visser, N. C. Voermans, L. B. O. Nijhuis et al.,“Quantification of trunk rotations during turning and walkingin Parkinson’s disease,” Clinical Neurophysiology, vol. 118, no.7, pp. 1602–1606, 2007.

[10] M. Ferrarin, I. Carpinella, M. Rabuffetti, E. Calabrese, P.Mazzoleni, and R. Nemni, “Locomotor disorders in patientsat early stages of parkinson’s disease: a quantitative analysis,”in Proceedings of the 28th Annual International Conference ofthe IEEE Engineering in Medicine and Biology Society (EMBS’06), pp. 1224–1227, New York, NY, USA, September 2006.

[11] G. M. Earhart and M. Hong, “Kinematics of podokineticafter-rotation: similarities to voluntary turning and potentialclinical implications,” Brain Research Bulletin, vol. 70, no. 1,pp. 15–21, 2006.

[12] C. R. Gordon, W. A. Fletcher, G. Melvill Jones, and E. W. Block,“Adaptive plasticity in the control of locomotor trajectory,”Experimental Brain Research, vol. 102, no. 3, pp. 540–545,1995.

[13] K. D. Weber, W. A. Fletcher, C. R. Gordon, G. MelvillJones, and E. W. Block, “Motor learning in the ’podokinetic’system and its role in spatial orientation during locomotion,”Experimental Brain Research, vol. 120, no. 3, pp. 377–385,1998.

[14] M. Hong, J. S. Perlmutter, and G. M. Earhart, “Podokineticafter-rotation in Parkinson disease,” Brain Research, vol. 1128,no. 1, pp. 99–106, 2007.

[15] M. Hong and G. M. Earhart, “Rotating treadmill trainingreduces freezing in Parkinson disease: preliminary observa-tions,” Parkinsonism and Related Disorders, vol. 14, no. 4, pp.359–363, 2008.

[16] B. A. Racette, M. Rundle, A. Parsian, and J. S. Perlmutter,“Evaluation of a screening questionnaire for genetic studies ofParkinson’s disease,” American Journal of Medical Genetics, vol.88, no. 5, pp. 539–543, 1999.

[17] C. G. Goetz, B. C. Tilley, S. R. Shaftman et al., “MovementDisorder Society-Sponsored Revision of the Unified Parkin-son’s Disease Rating Scale (MDS-UPDRS): scale presentationand clinimetric testing results,” Movement Disorders, vol. 23,no. 15, pp. 2129–2170, 2008.

[18] A. E. Patla, A. Adkin, and T. Ballard, “Online steering:coordination and control of body center of mass, head andbody reorientation,” Experimental Brain Research, vol. 129, no.4, pp. 629–634, 1999.

[19] J. R. Fuller, A. L. Adkin, and L. A. Vallis, “Strategies used byolder adults to change travel direction,” Gait and Posture, vol.25, no. 3, pp. 393–400, 2007.

[20] R. Grasso, C. Assaiante, P. Prevost, and A. Berthoz, “Devel-opment of anticipatory orienting strategies during locomotortasks in children,” Neuroscience and Biobehavioral Reviews, vol.22, no. 4, pp. 533–539, 1998.

[21] M. A. Hollands, A. E. Patla, and J. N. Vickers, “”Look whereyou’re going!”: gaze behaviour associated with maintainingand changing the direction of locomotion,” ExperimentalBrain Research, vol. 143, no. 2, pp. 221–230, 2002.

[22] M. A. Hollands, N. V. Ziavra, and A. M. Bronstein, “A newparadigm to investigate the roles of head and eye movementsin the coordination of whole-body movements,” Experimental

Brain Research, vol. 154, no. 2, pp. 261–266, 2004.[23] G. Courtine and M. Schieppati, “Human walking along a

curved path. I. Body trajectory, segment orientation and theeffect of vision,” European Journal of Neuroscience, vol. 18, no.1, pp. 177–190, 2003.

[24] S. B. Akram, J. S. Frank, and J. Fraser, “Effect of walkingvelocity on segment coordination during pre-planned turnsin healthy older adults,” Gait and Posture, vol. 32, no. 2, pp.211–214, 2010.

[25] M. Hong, J. S. Perlmutter, and G. M. Earhart, “A kinematicand electromyographic analysis of turning in people withParkinson disease,” Neurorehabilitation and Neural Repair, vol.23, no. 2, pp. 166–176, 2009.

[26] A. C. Lo, V. C. Chang, M. A. Gianfrancesco, J. H. Friedman,T. S. Patterson, and D. F. Benedicto, “Reduction of freezingof gait in Parkinson’s disease by repetitive robot-assistedtreadmill training: a pilot study,” Journal of NeuroEngineeringand Rehabilitation, vol. 7, no. 1, article 51, 2010.

[27] N. Giladi, D. McMahon, S. Przedborski et al., “Motor blocksin Parkinson’s disease,” Neurology, vol. 42, no. 2, pp. 333–339,1992.

[28] F. B. Horak, J. G. Nutt, and L. M. Nashner, “Postural inflex-ibility in parkinsonian subjects,” Journal of the NeurologicalSciences, vol. 111, no. 1, pp. 46–58, 1992.

[29] R. Inzelberg, M. Plotnik, T. Flash, E. Schechtman, I. Shahar,and A. D. Korczyn, “Mental and motor switching in Parkin-son’s disease,” Journal of Motor Behavior, vol. 33, no. 4, pp.377–385, 2001.

[30] N. Giladi, H. Shabtai, E. S. Simon, S. Biran, J. Tal, and A.D. Korczyn, “Construction of freezing of gait questionnairefor patients with Parkinsonism,” Parkinsonism and RelatedDisorders, vol. 6, no. 3, pp. 165–170, 2000.

[31] G. Frazzitta, R. Maestri, D. Uccellini, G. Bertotti, and P. Abelli,“Rehabilitation treatment of gait in patients with Parkinson’sdisease with freezing: a comparison between two physicaltherapy protocols using visual and auditory cues with orwithout treadmill training,” Movement Disorders, vol. 24, no.8, pp. 1139–1143, 2009.

[32] O. Bello, J. A. Sanchez, and M. Fernandez-del-Olmo, “Tread-mill walking in Parkinson’s disease patients: adaptation andgeneralization effect,” Movement Disorders, vol. 23, no. 9, pp.1243–1249, 2008.

[33] I. Miyai, Y. Fujimoto, Y. Ueda et al., “Treadmill trainingwith body weight support: its effect on Parkinson’s disease,”Archives of Physical Medicine and Rehabilitation, vol. 81, no. 7,pp. 849–852, 2000.



Research Article

Accuracy of Fall Prediction in Parkinson Disease:Six-Month and 12-Month Prospective Analyses

Ryan P. Duncan,1 Abigail L. Leddy,1 James T. Cavanaugh,2 Leland E. Dibble,3

Terry D. Ellis,4 Matthew P. Ford,5 K. Bo Foreman,3 and Gammon M. Earhart1, 6, 7

1 Program in Physical Therapy, Washington University in St. Louis School of Medicine, St. Louis, MO 63108, USA2 Department of Physical Therapy, University of New England, Portland, ME 04103, USA3 Department of Physical Therapy, University of Utah, Salt Lake City, UT 84108, USA4 Department of Physical Therapy and Athletic Training, Boston University, Boston, MA 02215, USA5 Department of Physical Therapy, University of Alabama at Birmingham School of Health Professions, Birmingham, AL 35294, USA6 Department of Anatomy & Neurobiology, Washington University in St. Louis School of Medicine, St. Louis, MO 63110, USA7 Department of Neurology, Washington University in St. Louis School of Medicine, St. Louis, MO 63110, USA


Received 26 July 2011; Revised 4 October 2011; Accepted 8 October 2011


Copyright © 2012 Ryan P. Duncan et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

Introduction. We analyzed the ability of four balance assessments to predict falls in people with Parkinson Disease (PD)prospectively over six and 12 months. Materials and Methods. The BESTest, Mini-BESTest, Functional Gait Assessment (FGA),and Berg Balance Scale (BBS) were administered to 80 participants with idiopathic PD at baseline. Falls were then tracked for12 months. Ability of each test to predict falls at six and 12 months was assessed using ROC curves and likelihood ratios (LR).Results. Twenty-seven percent of the sample had fallen at six months, and 32% of the sample had fallen at 12 months. At sixmonths, areas under the ROC curve (AUC) for the tests ranged from 0.8 (FGA) to 0.89 (BESTest) with LR+ of 3.4 (FGA) to 5.8(BESTest). At 12 months, AUCs ranged from 0.68 (BESTest, BBS) to 0.77 (Mini-BESTest) with LR+ of 1.8 (BESTest) to 2.4 (BBS,FGA). Discussion. The various balance tests were effective in predicting falls at six months. All tests were relatively ineffective at 12months. Conclusion. This pilot study suggests that people with PD should be assessed biannually for fall risk.

1. Introduction

Postural instability is a common cause of falls in peoplewith Parkinson disease (PD) [1]. In contrast to community-dwelling adults over age 65, approximately one-third ofwhom report falling each year [2], up to 70% of individualswith PD fall once annually, while 50% fall twice or more ina one year period [3, 4]. Falls lead to a myriad of com-plications [5] that can affect not only physical health, butalso the psychological health of the individual. Hip fractureand head trauma are two of the most common physicalproblems incurred by an individual with PD following a fall[6], while the psychological complications include fear offalling [7, 8] and reduced quality of life [9]. Such fall-relatedcomplications are associated with substantial economic costs

[10, 11] and indicate an urgent need to identify and protectthose individuals at the greatest risk.

Despite the relatively high prevalence of falls in the PDpopulation, accurate and useful methods for predicting animpending future fall, especially during the early stages ofthe disease, remain elusive. Fall history, a well-known fallrisk factor among older adults [12], has a limited utility asa solitary predictive indicator. Although a meta-analysis ofprospective studies of falling in PD found that 57% of indi-viduals who had a history of falls in the past year fell duringa 3-month surveillance period, so did 21% of individualswith no history of falls [13]. Moreover, fall incidence alonedoes not help to identify underlying contributors to posturalinstability specific to PD. People with PD, for example, maydemonstrate impairments in areas of movement control such


as sensory integration, keeping their center of mass withintheir base of support, coordination of anticipatory posturalcontrol tasks [8, 14] as well as medication side effects suchas dyskinesias [15]. For this reason, standardized balanceassessment tools have been recommended to help determinefactors contributing to falls so that therapeutic interventiontargets can be identified [16, 17].

The utility of a variety of clinical balance tests has beenstudied. Balance assessments including the Tinetti [18], BergBalance Scale (BBS) [19], the Timed Up and Go (TUG) [20],the Functional Gait Assessment (FGA) [21], and recentlydeveloped Balance Evaluation Systems Test (BESTest) [22]have been shown to have sensitivity and specificity thatexceeds a random guess, but they still demonstrate a clini-cally relevant proportion of false-positive and false- negativepredictions [5, 23]. As noted in a previous meta-analysis[13], new prediction methods are needed. Relatively newlydeveloped balance assessments such as the Functional GaitAssessment (FGA) [21], BESTest [22], and Mini-BESTest, acondensed version of the BESTest [24], have yet to be studiedand compared prospectively.

Regardless of the balance assessment utilized, there havebeen efforts to improve the predictive performance on thesebalance assessments through diagnosis-specific alterations ofcutoff scores or collective interpretation of multiple tests [15,23, 25, 26]. While these methods may improve accuracy, theiroverall success may be limited by participant’s fall recall bias.To date, we are unaware of any studies that have examinedand compared whether the length of prospective follow-upaffects the accuracy of fall prediction in persons with PD.

In order to address these gaps in our understanding of fallprediction in persons with PD, the primary objective of thisstudy was to compare the relative accuracy for fall predictionof four common balance assessments at the six-month and12-month prospective time points. Relative to our primaryobjective, we hypothesized that these tests would be useful inpredicting falls prospectively at both six and 12 months, withbetter accuracy over the shortest of the two time periods. Oursecondary objective was to compare the predictive accuracyand the validity indices of the four balance assessments.Relative to our secondary objective, we hypothesized thattests such as the FGA, BESTest, and Mini-BESTest that in-corporate dynamic tasks would demonstrate improved pre-dictive ability compared to the BBS.

2. Methods

2.1. Participants. We recruited participants using contact in-formation gathered from the Washington University Schoolof Medicine’s Movement Disorders Center database and theVolunteers for Health database. Participants were recruitedas part of a larger study [27]. Individuals were included ifthey had a medical diagnosis of idiopathic PD (Hoehn andYahr (H&Y) Stages I–IV), were over the age of 40 and werecommunity dwellers. Study candidates were excluded if theyhad atypical parkinsonism or previous surgical managementof PD (pallidotomy or deep brain stimulation). Prior toparticipation, each participant provided written informedconsent in accordance with the policies and procedures of

Washington University School of Medicine’s Human Re-search Protection Office.

2.2. Data Collection. Participants were evaluated at baselineutilizing four balance tests (BBS, FGA, BESTest, and Mini-BESTest) as described below under Balance Assessments.Participants were then followed for 12 months, with fallincidence determined through participant’s report at thesix-month and 12-month time points. An individual wasconsidered a faller if he or she reported two or more fallsover the surveillance period of interest (0–6 months or 0–12months). An individual was considered a nonfaller if he orshe reported zero or one falls during the surveillance period.

2.3. Balance Assessments. The BBS is a well-established bal-ance measure consisting of 14 items (sit to stand, transfers,forward reach, etc.) used to determine whether or not onemay be at risk for falls [28]. The BBS does not evaluate thebalance during walking. It has been shown to be reliablewhen used to assess balance in people with PD [29]. Eachitem is scored on a scale of zero (indicating impairedbalance) to four (indicating no impairment in balance), witha maximum possible score of 56.

The FGA [21] is a 10-item test of dynamic balance inwhich all components are evaluated while the participant iswalking. Items performed by the participant include forwardand backward walking as well as walking while turningthe head, changing walking speeds, stepping over obstacles,and walking with a narrow base of support. When used toevaluate individuals with PD, this test had high interrater andtest-retest reliability [23]. Each item is scored on a scale ofzero (indicating loss of balance, increased time to performtask, significantly altered gait pattern) to three (indicating noimpairment of gait or balance and completion of the task ina timely manner), with a maximum possible score of 30.

The BESTest [22] is a measure designed to evaluate bal-ance control via 36 items that are divided into six sections(biomechanical constraints, stability limits and verticality,anticipatory postural adjustments, postural responses, sen-sory orientation, and stability in gait). Items in the BESTestinclude selected items from the aforementioned assessments(i.e., BBS and FGA) as well as items such as center ofmass alignment, hip and ankle strength, sitting verticalityand lateral lean, and multidirectional compensatory steppingcorrection, among others. The BESTest has high interraterand test-retest reliability in PD [23]. Each item is scored on ascale of zero (indicating poor balance or inability to completetask) to three (no impairment in balance), with a maximumscore of 108 points.

A shortened version of the BESTest, the Mini-BESTest,was designed “to improve the structure and measurementqualities” of the BESTest [24]. This shorter version can beadministered more quickly than the full BESTest, therebyreducing clinician and patient burden. The Mini-BESTest isa 14-item balance evaluation that concentrates on dynamicbalance and its components are derived from four of thesix BESTest sections. Items are scored on a scale of zero(poor balance) to two (no impairment of balance), with


a maximum possible score of 32 as two of the 14 items receivetwo separate scores for different aspects of the tasks [30].

2.4. Procedures. All balance assessments were administeredin the Locomotor Control Laboratory at Washington Uni-versity School of Medicine by a trained physical therapist.Baseline assessments of participants began in July and endedin December of 2009. All participants maintained theirnormal medication regimen so that they were tested in the“on” phase of their medication, one to two hours aftermedication intake. Demographic information, fall incidence,and Movement Disorder Society Unified Parkinson DiseaseRating Scale Motor Subscale III (MDS-UPDRS-III) scoreswere obtained prior to administration of balance assessments[31, 32]. Regarding fall incidence, participants were followedprospectively and at six months reported how many timesthey fell in the period from baseline to six months. At 12months, participants reported how many times they fellin the period between six and 12 months, with numberof falls from 0 to 12 months determined by adding thetwo reports together. Participants chose from the followinganswers: (1) none, (2) one time, (3) 2–10 times, (4) weekly,or (5) daily. An individual was classified as a faller if heexperienced two or more falls in the period of interest (i.e.,from baseline to six months or from baseline to 12 months).The order of balance assessments was as follows: BBS, FGA,and BESTest. Mini-BESTest scores were derived from theBESTest item scores, as all items on the Mini-BESTest areincluded in the BESTest. Items that were duplicated amongthe BBS, FGA, and BESTest were performed only once andscored appropriately for each tool. For example, a sit-to-stand transfer task is in the BBS and BESTest; therefore it wasonly performed once and scored by the rater according to thecriteria listed on each tool.

2.5. Data Analysis. In order to test our primary hypoth-esis, receiver operating characteristic curves (ROCs) wereconstructed for each balance assessment at each time point(six and 12 months) and the area under the curve (AUC)was determined for each test at each time point. Usingpreviously established cutoff scores [23], we determined thearea under the curve (AUC), positive and negative likelihoodratios, and posttest probabilities for each test at each timepoint [33–35]. The time point that consistently producedthe balance assessments with higher AUC and positive LR aswell as lower negative LR would be interpreted as the moreaccurate time point. Once that determination was made, weexamined our secondary objective and hypothesis throughthe use of empirical tests for noninferiority that were used tomake pairwise comparisons of the AUC for each test (P <0.05) [36]. Point estimators and interval estimators (95%confidence intervals [95% CI]) were calculated for all AUCand likelihood ratio values.

3. Results

Baseline evaluations were completed on 80 participants.Of the original cohort, 51 participants (41% male) com-pleted the six-month evaluation, and 40 participants (40%

Table 1: Demographics.

6-Month Group(n = 51)

12-Month Group(n = 40)

Age 67.5± 8.8 67.3± 9.5

Years with diagnosis 7.7± 3.9 7.2± 4.1

UPDRS motor score 39.3± 13.3 37.8± 13.1

H&Y stages 2.4± 0.6 2.3± 0.6

Fallers (pretestprobability of falling)

14/15 (0.275) 13/40 (0.325)

male) completed the 12-month evaluation (Table 1). Atsix months, 14 individuals (27%) were considered fallers,while 13 individuals (32%) were considered fallers at 12months. Regarding reasons for dropout at six months, 15participants were unable to be contacted or gave no reasonfor discontinuing, nine experienced a decline in conditionor an unrelated medical condition, one had transportationdifficulty, one participant experienced family problems, andthree participants had incomplete data sets. At 12 months,in addition to those who had dropped out by six months,four participants were unable to be contacted or gave noreason for discontinuing, three experienced a decline incondition or an unrelated medical condition, and fourparticipants had incomplete data sets. Of the 11 individualsthat were lost from six to 12 months, seven (three males) werecharacterized as fallers at six months.

3.1. Comparison of Six- and 12-Month Results. At six months,AUCs for the tests ranged from 0.8 to 0.89, while at 12months, AUCs ranged from 0.68 to 0.77. At six months(Table 2(a)), the positive likelihood ratios were greater,the negative likelihood ratios were lower, and the posttestprobability values were lower (i.e., better) for all for balancetests than at 12 months (Table 2(b)).

3.2. Individual Test Comparison. Based on the apparentgreater accuracy of the six-month prediction, the individualtests were compared at the six-month time point to deter-mine which, if any, was superior to the others in terms ofpredictive ability. All tests provided greater accuracy thana random guess, with AUC point estimators ranging from0.89 (BESTest) to 0.80 (FGA) and substantially overlapping95% CIs (Table 2(a), Figure 1). However, noninferiority testsrevealed that the AUC of the BESTest was superior to that ofall other tests. Noninferiority tests also showed that the FGAwas inferior to all other tests (Table 3).

4. Discussion

Previous prospective studies of fall prediction have utilizedvaried lengths of follow-up period [5, 13]. However, toour knowledge, no previous work has directly comparedthe accuracy of fall prediction at different follow-up inter-vals. Our data confirmed our primary hypothesis that ashorter follow-up period (six months) consistently producedmore accurate predictions than a longer follow-up period


Table 2

(a) Predictive values at 6 months.

Balancemeasure

AUC (95% CI) Score Sensitivity Specificity LR + (95% CI) LR− (95% CI)

Posttestprobabilitywith test ≤cutoff value

Posttestprobabilitywith test >cutoff value

BESTest 0.89 (0.74–0.95) ≤69% 0.93 0.84 5.81 (3.69–9.14) 0.08 (0.04–0.17) 0.69 0.03

Mini-BESTest

0.87 (0.72–0.94) ≤20/32(63%) 0.86 0.78 3.97 (2.68–5.70) 0.18 (0.11–0.78) 0.60 0.07

BBS 0.87 (0.75–0.95) ≤47/56 0.79 0.86 5.64 (3.43–9.27) 0.24 (0.17–0.36) 0.68 0.09

FGA 0.80 (0.62–0.90) ≤15/30 0.64 0.81 3.37 (2.19–5.18) 0.44 (0.34–0.59) 0.56 0.15

(b) Predictive values at 12 months.

Balancemeasure

AUC (95% CI) Score Sensitivity Specificity LR + (95% CI) LR− (95% CI)

Posttestprobabilitywith test ≤cutoff value

Posttestprobabilitywith test >cutoff value

BESTest 0.68 (0.45–0.83) ≤69% 0.46 0.74 1.77 (1.19–2.62) 0.73 (0.59–0.91) 0.46 0.26

Mini-BESTest

0.77 (0.55–0.89) ≤20/32(63%) 0.62 0.74 2.37 (1.66–3.34) 0.52 (0.39–0.68) 0.53 0.20

BBS 0.68 (0.45–0.82) ≤47/56 0.46 0.81 2.42 (1.53–3.82) 0.67 (0.54–0.82) 0.54 0.24

FGA 0.70 (0.50–0.83) ≤15/30 0.46 0.81 2.42 (1.53–3.82) 0.67 (0.54–0.82) 0.54 0.24

(12 months). In addition, at the six-month follow-up timepoint, all of the balance assessments studied provided clin-ically useful predictive accuracy. Comparisons of the pointestimators and statistical tests of noninferiority suggestedthat the BESTest produced the greatest predictive accuracy.However, it is unclear whether the differences between theBESTest and the other balance measures are sufficiently largeto merit use of one test over another in a clinical setting.

4.1. When Should Fall-Related Screening Take Place? Therecently published American Academy of Neurology qualityof care measures for Parkinson Disease state that personswith PD should be assessed for fall-related issues “at leastannually [37].” While these guidelines provide targets forclinicians, they were developed through a consensus buildingprocess that involved expert panel input, public comment,and stakeholder input, and therefore lacked research-basedsupport. Our findings of both six- and 12-month predictiveaccuracy having AUC values greater than 0.50 (the level ofa random guess) support this metric. However, if clinicianswish to most accurately assess the risk of falling of a personwith PD, our data suggest that they should consider thatbiannual follow-up of persons with PD regarding falls.

4.2. Is One Test Better Than Another? The validity indices(AUC, positive and negative likelihood ratios) demonstratedthat all of the tests studied provided clinically meaningfulpredictive ability. Substantial overlap of the interval esti-mators agreed with previous studies that have documentedmoderate levels of accuracy for the BBS and the FGA [16, 23].In this sample, the point estimators of the validity indicesand the tests for noninferiority indicated that the BESTest

provided the highest level of accuracy and, for the firsttime, provided prospective documentation of its predictivevalidity. The BESTest’s likelihood ratio modifications to thepretest probability of being a faller provide a specific exampleof the clinical relevance of these findings. At six months, thepretest probability of being a faller was 27%. Based on theBESTest positive likelihood ratio, an individual who scoredbelow the cutoffs for the BESTest increased their posttestprobability of being a faller to 69%. Based on the BESTestnegative likelihood ratio, an individual with a score above thecutoff reduced their posttest probability of being a faller to3%. These modifications to the pretest probability are similarto those observed in other studies of persons with PD [25].

While our results suggested that the BESTest may bethe most accurate as a free-standing test to predict falls inthe absence of other balance assessments, the administrationtime of the BESTest is much longer than the other threeassessments. Although the results of this study support itsuse when assessing balance and fall risk in individuals withPD, it is not clear whether the slightly improved accuracy ofthe BESTest as compared to the Mini-BESTest or BBS at sixmonths is enough to merit utilization of the full BESTest inclinical settings where time constraints must be considered.

We found it surprising that the BESTest, Mini-BESTest,and BBS outperformed the FGA when used prospectivelyover six months. Based on previous research, we hypothe-sized that more dynamic balance tests such as the BESTest,Mini-BESTest, and FGA would be more likely to accuratelypredict falls than a less dynamic balance test like the BBS[23, 38]. However, our findings regarding this hypothesiswere mixed, with the FGA having the lowest predictiveaccuracy in this sample. Regardless of the FGA findings, our


BESTestBBSFGA

0 0.2 0.4 0.6 0.8 1

Sen

siti

vity

0

0.2

0.4

0.6

0.8

1

1-specificity

Mini-BESTest

Six-month ROC curves

(a)

BESTestBBSFGA

0 0.2 0.4 0.6 0.8 1

Sen

siti

vity

0

0.2

0.4

0.6

0.8

1

1-specificity

Mini-BESTest

Twelve-month ROC curves

(b)

Figure 1

Table 3: P Values for 6 month NonInferiority test of AUC.

BESTest Mini-BESTest BBS FGA

BESTest — 0.02 0.04 <0.001

Mini-BESTest 0.13 — 0.14 0.001

BBS 0.23 0.19 — 0.01

FGA 0.83 0.73 0.65 —

First variable in comparison is listed vertically, second variable is listedhorizontally. Bold values indicate significant difference where first variableis superior to second variable.

results generally agreed with recent research advocating forecologically valid balance assessments [16].

4.3. Limitations and Directions for Future Research. Whileour results suggest that balance assessments may be justifiedon a biannual basis, these results should be interpreted

with some caution. First, our sample size for this pilotstudy was small and representative of a cohort with onlymild-to-moderate PD severity with a smaller percentage offallers than seen in previous balance assessment validitystudies. In addition, a moderate number of participants werelost to follow-up at the six- and 12-month measurementpoints. Future research should examine larger samples ofparticipants over a broader spectrum of disease severity andperhaps also consider different motor phenotypes within PD.

Second, we utilized previously established cut-off scoresfor all of the four balance assessments. These cut-off scoresstill resulted in false-negative and false-positive predictions.Since cut-off scores based on validity indices will likelychange depending on the sample being studied, it is impor-tant to emphasize that cutoff scores should be utilized withcaution and with the appreciation that any and all cutoffscores are simply guidelines and not definitive boundariesthat separate fallers from nonfallers.

Third, our method of collecting fall incidence data, whenused over a period of six months or more, can lead to anunderreporting of falls [39]. As such, we suggest that futurestudies follow more rigorous procedures for collecting fallincidence data as outlined by Lamb and colleagues [40].Future studies may also be designed to assess people with PDoff anti-Parkinson medication to determine whether falls aremore likely during this state.

5. Conclusion

Prospective identification of fall risk for individuals withPD is extremely important in order to demonstrate a needfor therapeutic intervention aimed at reducing fall risk.Our comparison of varied duration of follow-up revealedthat a six-month follow-up resulted in greater accuracyof fall prediction than a 12-month follow-up. In terms ofaccuracy of fall prediction during that six-month follow-upperiod, all tests provided moderate-to-strong accuracy forfall prediction with clinically meaningful alterations in theprobability of being a faller. While the BESTest was slightlymore accurate than the other tests, no test eliminated false-positive and false-negative predictions.

5.1. Rehabilitation Implications. None of the tests examinedpossesses acceptable predictive ability in determining who isat risk for falls within the next 12 months, suggesting theneed for regular balance evaluations every six months amongpeople with PD. Such a model of preventative evaluationand treatment twice per year is in keeping with othermodels of healthcare, such as the well-established system ofprophylactic dental care in the United States. Such a modelwould likely be appropriate and beneficial to apply in therehabilitative care of individuals with PD.

Acknowledgments

The authors thank the Davis Phinney Foundation and theParkinson’s Disease Foundation for their financial support ofthis study. Sincere thanks are due to the participants for theirinvolvement in this study.


References

[1] M. Rudzinska, S. Bukowczan, K. Banaszkiewicz, J. Stozek, K.Zajdel, and A. Szczudlik, “Causes and risk factors of falls inpatients with Parkinson’s disease,” Neurologia i NeurochirurgiaPolska, vol. 42, no. 3, pp. 216–222, 2008.

[2] L. Z. Rubenstein and K. R. Josephson, “Falls and their preven-tion in elderly people: what does the evidence show?” MedicalClinics of North America, vol. 90, no. 5, pp. 807–824, 2006.

[3] B. H. Wood, J. A. Bilclough, A. Bowron, and R. W. Walker,“Incidence and prediction of falls in Parkinson’s disease: aprospective multidisciplinary study,” Journal of NeurologyNeurosurgery and Psychiatry, vol. 72, no. 6, pp. 721–725, 2002.

[4] Y. A. M. Grimbergen, M. Munneke, and B. R. Bloem, “Fallsin Parkinson’s disease,” Current Opinion in Neurology, vol. 17,no. 4, pp. 405–415, 2004.

[5] G. K. Kerr, C. J. Worringham, M. H. Cole, P. F. Lacherez, J. M.Wood, and P. A. Silburn, “Predictors of future falls in Parkin-son disease,” Neurology, vol. 75, no. 2, pp. 116–124, 2010.

[6] L. J. Melton, C. L. Leibson, S. J. Achenbach et al., “Fracturerisk after the diagnosis of Parkinson’s disease: influence ofconcomitant dementia,” Movement Disorders, vol. 21, no. 9,pp. 1361–1367, 2006.


[8] A. L. Adkin, J. S. Frank, and M. S. Jog, “Fear of falling andpostural control in Parkinson’s disease,” Movement Disorders,vol. 18, no. 5, pp. 496–502, 2003.

[9] D. H. Romero and G. E. Stelmach, “Changes in postural con-trol with aging and Parkinson’s disease,” IEEE Engineering inMedicine and Biology Magazine, vol. 22, no. 2, pp. 27–31, 2003.

[10] W. E. Bacon, “Secular trends in hip fracture occurrence andsurvival: age and sex differences,” Journal of Aging and Health,vol. 8, no. 4, pp. 538–553, 1996.

[11] J. A. Stevens, P. S. Corso, E. A. Finkelstein, and T. R. Miller,“The costs of fatal and non-fatal falls among older adults,”Injury Prevention, vol. 12, no. 5, pp. 290–295, 2006.

[12] J. T. Chang and D. A. Ganz, “Quality indicators for falls andmobility problems in vulnerable elders,” Journal of the Ameri-can Geriatrics Society, vol. 55, no. 2, pp. S327–S334, 2007.

[13] R. M. Pickering, Y. A. M. Grimbergen, U. Rigney et al.,“A meta-analysis of six prospective studies of falling inParkinson’s disease,” Movement Disorders, vol. 22, no. 13, pp.1892–1900, 2007.



[16] K. B. Foreman, O. Addison, H. S. Kim, and L. E. Dibble, “Test-ing balance and fall risk in persons with Parkinson disease,an argument for ecologically valid testing,” Parkinsonism andRelated Disorders, 2011.

[17] M. D. Latt, S. R. Lord, J. G. L. Morris, and V. S. C. Fung,“Clinical and physiological assessments for elucidating fallsrisk in Parkinson’s disease,” Movement Disorders, vol. 24, no. 9,pp. 1280–1289, 2009.

[18] M. E. Tinetti, “Performance-orientated assessment of mo-bility problems in elderly patients,” Journal of the AmericanGeriatrics Society, vol. 34, no. 2, pp. 119–126, 1986.

[19] K. Berg, S. Wood-Dauphinee, J. I. Williams, and D. Gayton,“Measuring balance in the elderly: preliminary developmentof an instrument,” Physiotherapy Canada, vol. 41, no. 6, pp.304–311, 1989.

[20] D. Podsiadlo and S. Richardson, “The timed “Up and Go”:a test of basic functional mobility for frail elderly persons,”Journal of the American Geriatrics Society, vol. 39, no. 2, pp.142–148, 1991.

[21] D. M. Wrisley, G. F. Marchetti, D. K. Kuharsky, and S. L.Whitney, “Reliability, internal consistency, and validity ofdata obtained with the functional gait assessment,” PhysicalTherapy, vol. 84, no. 10, pp. 906–918, 2004.

[22] F. B. Horak, D. M. Wrisley, and J. Frank, “The balanceevaluation systems test (BESTest) to differentiate balancedeficits,” Physical Therapy, vol. 89, no. 5, pp. 484–498, 2009.

[23] A. L. Leddy, B. E. Crowner, and G. M. Earhart, “Functionalgait assessment and balance evaluation system test: reliability,validity, sensitivity, and specificity for identifying individualswith parkinson disease who fall,” Physical Therapy, vol. 91, no.1, pp. 102–113, 2011.

[24] F. Franchignoni, F. Horak, M. Godi, A. Nardone, and A.Giordano, “Using psychometric techniques to improve thebalance evaluation systems test: the mini-bestest,” Journal ofRehabilitation Medicine, vol. 42, no. 4, pp. 323–331, 2010.

[25] L. E. Dibble and M. Lange, “Predicting falls in individualswith Parkinson disease: a reconsideration of clinical balancemeasures,” Journal of Neurologic Physical Therapy, vol. 30, no.2, pp. 60–67, 2006.

[26] M. R. Landers, A. Backlund, J. Davenport, J. Fortune, S.Schuerman, and P. Altenburger, “Postural instability in idio-pathic parkinson’s disease: discriminating fallers from non-fallers based on standardized clinical measures,” Journal ofNeurologic Physical Therapy, vol. 32, no. 2, pp. 56–61, 2008.

[27] L. E. Dibble, J. T. Cavanaugh, G. M. Earhart, T. D. Ellis, M. P.Ford, and K. B. Foreman, “Charting the progression of dis-ability in parkinson disease: study protocol for a prospectivelongitudinal cohort study,” BMC Neurology, vol. 10, article110, 2010.

[28] K. O. Berg, S. L. Wood-Dauphinee, J. I. Williams, and B. Maki,“Measuring balance in the elderly: validation of an instru-ment,” Canadian Journal of Public Health, vol. 83, no. 2, pp.S7–S11, 1992.

[29] L. I. I. K. Lim, E. E. H. Van Wegen, C. J. T. De Goede etal., “Measuring gait and gait-related activities in Parkinson’spatients own home environment: a reliability, responsivenessand feasibility study,” Parkinsonism and Related Disorders, vol.11, no. 1, pp. 19–24, 2005.

[30] A. L. Leddy, B. E. Crowner, and G. M. Earhart, “Utility ofthe mini-BESTest, BESTest, and BESTest sections for balanceassessments in individuals with Parkinson disease,” Journal ofNeurologic Physical Therapy, vol. 35, no. 2, pp. 90–97, 2011.

[31] C. G. Goetz, S. Fahn, P. Martinez-Martin et al., “Movementdisorder society-sponsored revision of the unified Parkinson’sdisease rating scale (MDS-UPDRS): process, format, andclinimetric testing plan,” Movement Disorders, vol. 22, no. 1,pp. 41–47, 2007.

[32] C. G. Goetz, B. C. Tilley, S. R. Shaftman et al., “Movementdisorder society-sponsored revision of the unified Parkinson’sdisease rating scale (MDS-UPDRS): scale presentation andclinimetric testing results,” Movement Disorders, vol. 23, no.15, pp. 2129–2170, 2008.

[33] A. K. Akobeng, “Understanding diagnostic tests 3: receiveroperating characteristic curves,” Acta Paediatrica, vol. 96, no.5, pp. 644–647, 2007.


[34] J. J. Deeks and D. G. Altman, “Diagnostic tests 4: likelihoodratios,” British Medical Journal, vol. 329, no. 7458, pp.168–169, 2004.

[35] M. Greiner, D. Pfeiffer, and R. D. Smith, “Principles andpractical application of the receiver-operating characteristicanalysis for diagnostic tests,” Preventive Veterinary Medicine,vol. 45, no. 1-2, pp. 23–41, 2000.

[36] J. Hintze, NCSS, NCSS, LLC, Kaysville, Utah, 2009.[37] E. M. Cheng, S. Tonn, R. Swain-Eng, S. A. Factor, W. J. Weiner,

and C. T. Bever, “Quality improvement in neurology: AANParkinson disease quality measures: report of the QualityMeasurement and Reporting Subcommittee of the AmericanAcademy of Neurology,” Neurology, vol. 75, no. 22, pp.2021–2027, 2010.

[38] A. C. Dennison, J. V. Noorigian, K. M. Robinson et al.,“Falling in Parkinson disease: identifying and prioritizingrisk factors in recurrent fallers,” American Journal of PhysicalMedicine and Rehabilitation, vol. 86, no. 8, pp. 621–632, 2007.

[39] S. R. Cummings, M. C. Nevitt, and S. Kidd, “Forgetting falls.The limited accuracy of recall of falls in the elderly,” Journalof the American Geriatrics Society, vol. 36, no. 7, pp. 613–616,1988.

[40] S. E. Lamb, E. C. Jørstad-Stein, K. Hauer, and C. Becker,“Development of a common outcome data set for fall injuryprevention trials: the Prevention of Falls Network Europeconsensus,” Journal of the American Geriatrics Society, vol. 53,no. 9, pp. 1618–1622, 2005.


Research Article

Community Walking in People with Parkinson’s Disease

Robyn M. Lamont,1 Meg E. Morris,2 Marjorie H. Woollacott,3 and Sandra G. Brauer1

1 School of Health and Rehabilitation Sciences, The University of Queensland, Brisbane, QLD 4072, Australia2 Melbourne School of Health Sciences, The University of Melbourne, Melbourne, VIC 3010, Australia3 Department of Human Physiology, University of Oregon, Eugene, OR 97401, USA

Correspondence should be addressed to Robyn M. Lamont, [email protected]



Copyright © 2012 Robyn M. Lamont et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

People with Parkinson’s disease often have walking difficulty, and this is likely to be exacerbated while walking in places in thecommunity, where people are likely to face greater and more varied challenges. This study aims to understand the facilitators andthe barriers to walking in the community perceived by people with Parkinson’s disease. This qualitative study involved 5 focusgroups (n = 34) of people with Parkinson’s disease and their partners residing in metropolitan and rural regions in Queensland,Australia. Results found that people with PD reported to use internal personal strategies as facilitators to community walking, butidentified primarily external factors, particularly the environmental factors as barriers. The adoption of strategies or the use offacilitators allows people with Parkinson’s disease to cope so that participants often did not report disability.

1. Introduction

Community ambulation is compromised in many peopleliving with Parkinson’s disease (PD), which is thought toaffect around 2 percent of the population over the age of65 [1]. Gait changes are a hallmark of PD, and people withPD frequently walk with reduced speed and step length[2, 3], reduced cadence [2–5], and increased gait variability[6]. People with PD may also experience freezing whenwalking. Walking difficulties are exacerbated when attentionis drawn away from walking by performing additional tasks[5–9]. Challenging environments that demand attention mayalso compromise the ability to walk in people with thisdebilitating condition.

Community walking is an important enabler to par-ticipation in community activities and a range of societal,work, and leisure roles. It has been defined as locomotionin environments outside the home or the residence [10].This includes the ability to negotiate public and privatevenues both indoors and outdoors that incorporate a varietyof environmental demands [10, 11], which could provechallenging for people with PD.

The physical, social, and attitudinal environments aregenerally more varied and less predictable in the community

than for the home or the laboratory settings. Walking inthe community is generally assumed to be a more complexand high-level skill than walking around the home or inthe laboratory. Research in older adults suggests that lossof walking function is a gradual process which results ina restriction of the variety of places they go to and thedistance they will venture from home [12]. Impairmentscan accelerate this, and disabled older adults report fewerencounters with and greater avoidance of physical challengesin the environment [13].

People living with PD have walking challenges in addi-tion to the usual ageing process. The impact of thesechallenges on community walking is not yet understood. Agreater understanding of the perceived factors (both internaland external to the person) that positively and negativelyimpact on the ability of people with PD to walk in the com-munity is needed. Understanding these factors may allowclinicians to design assessment tools more appropriate formeasuring community mobility deficits and provide a basisfor the development of interventions to improve communitymobility and potentially participation in people with PD.The aim of this qualitative study is to understand whatspecific facilitators and barriers individuals with PD perceiveaffect their ability to walk successfully in the community.


Table 1: Demographic information of study participants.

Participant number PD/partner Age (yrs) Disease duration (yrs) Freezing of gait Falls in past 6 mths Group type

1 PD 62 12 No 5 Metro

2 PD 75 21 Yes 16 Metro




6 PD 71 19 No Daily Metro




10 PA — — — — Metro

11 PA — — — — Metro


13 PA 57 — — — Metro


15 PA 64 — — — Metro

16 PD 54 6 No 0 Rural

17 PD 41 5 Yes 0 Rural

18 PA 39 — — — Rural


20 PA — — — — Rural


22 PA 70 — — — Rural

23 PA 76 — — — Metro



26 PA 76 — — — Metro


28 PA 60 — — — Metro

29 PA 76 — — — Partner






Mean age 67 years, range 41–82 years.Mean disease duration 10 years, range 4–21 years.

2. Methods

A qualitative study design was used to allow data to begathered directly from people living with PD. Focus groupswere used with the aim of encouraging discussion of a varietyof experiences and opinions. Data collection ceased whensaturation of the data was achieved.

2.1. Participants. People with PD and partners of people withPD were recruited using advertising in local PD Associationpublications in Queensland, Australia. Participants wereeligible for the study if they or their partner had PD orthey cared for someone with PD, were able to sign informedconsent, and able to attend a focus group in a communitysetting.

Five focus groups were conducted (n = 34) includingthree metropolitan groups of people with PD and theirpartners (n = 22), one metropolitan group of partnersonly (n = 6), and one rural group (n = 7). A partnergroup was included as it was felt that partners of peoplewith PD could have a valuable contribution to make to thisdata collection but that some may be reluctant to honestlyexpress their feelings regarding the ability of their partnerif they were present. The group of partners of people withPD was purposively sampled using a database of peoplewilling to participate in research related to PD. Demographicinformation about the participants is included in Table 1.

2.2. Procedure. Each focus group included the participants,a facilitator, and a scribe who took field notes regarding


Table 2: Key focus group questions.

(1) Why do you walk outside your home?

(2) How is walking in the community different to walking at home?

(3) What factors make walking in the community easier?

(4) What factors make walking in the community difficult?

group dynamics, nonverbal communication, and interview-ing conditions. Groups lasted one to two hours and wereaudio recorded. Prior to each focus group, participants weregiven written information outlining the aim of the research,the procedure for the session, and an outline of the 4 keyquestions (see Table 2) for discussion. They were given theopportunity to ask any questions, provided written informedconsent, and completed a short questionnaire of generaldemographic information.

Key questions were open ended so responses were inparticipants’ own words. Probing questions were used whenneeded, but every effort was made to maintain a naturaldiscussion. At the end of each focus group, the facilitatorsummarized the main points of the discussion and herperceptions. Participants were asked to confirm the accuracyof this summary.

Approval for this study was obtained from the Universityof Queensland’s Behavioural and Social Sciences EthicalReview Committee (Application #2008001843).

2.3. Analysis. Immediately after each group, the facilitatorreflected on the discussion with the aim of putting asideany immediate thoughts or judgments so the next group wasapproached with minimal preconceptions.

All audio recordings were professionally transcribedverbatim by professionals external to the study. To confirmaccuracy, members of the research team checked eachtranscription twice against the audio file. Two researchers(RL & SB) then performed thematic content analysis ofthe transcripts, using a process of repeated readings. Initialreading aimed to capture the context of the entire discus-sion. Further readings aimed to identify themes that wereemerging with notes initially made in the margins identifyingnoteworthy phrases, lines, and paragraphs of the prose.These were analysed, asking first “what does this mean?” andthen “how is this the same/different to other segments?” [14].At this point the two researchers met to discuss the themeseach had identified and classify the distinctive features ofthese themes. Subsequent readings of the transcripts wereperformed to ensure the accuracy of the themes and toidentify sections of discussion consistent and inconsistentwith these themes.

At this point the researchers performed an analysis ofthe existing literature. This ensured that themes were drawnsolely from the data without influence of preconceived ideasinterpreted from the literature.

3. Results

Eighteen people with PD with a mean age of 67 years (range41–82 years) and mean disease duration of 10.3 years (range

4–21 years) participated in the study. Freezing was reportedby 44% (8) of participants, and 33% (6) reported falls in theprior 6 months (Table 1). Twenty-two partners who had amean age of 65.4 (range 39–78) were also included.

Three primary themes emerged from the data: (i) peoplewith PD used internal and external facilitators to makewalking in the community easier, (ii) they perceived barriersto be primarily external environmental factors, and (iii) dueto their effective use to/of facilitatory strategies, many peoplewith PD did not report community walking disability. Thesewill be outlined in turn.

3.1. Facilitators. Several factors which contribute to theability of a person to walk in the community were discussedby the groups. These are termed facilitators and includedboth internal factors driven by the person and externalfactors mediated by objects or people outside the person withPD. Internal factors were often strategies people adoptedto ensure they could continue to optimally walk in thecommunity. These could be spontaneous strategies, used tocope with a particular situation or symptom as it arose,planned in advance to maximise the chance of success, ormay have become a normal behaviour now used withoutcompromise.

3.1.1. Internal Facilitators. A common strategy described wasconsciously attending to walking speed, step length, and toeclearance. This strategy was reported to be used to respondto challenges to walking when they arose. Most people whoreported gait changes described using this strategy as eitherconcentrating on their walking or taking extra care withwalking.

“But if you walk slower and lift your feet andconcentrate that helps” (PD-27).

While thinking of taking long, rhythmical steps wascommonly used to aid walking in the community, it wasreported that remembering to use this strategy in a commu-nity environment may be less automatic than when at home.

“. . .you’ve got to try and think and remember todo it, like, think and make sure you do it . . . tryand step it out and lift your feet more” (PD-27).

Planning and preparation played a role to ensure walkingin the community was successful. Almost everyone reportedtiming outings to coincide with times of high medicationeffectiveness (“ON” times). Being prepared for outings,making a plan and keeping to that plan reduced the chanceof running late, feeling rushed, and making errors such asforgetting to take medications, and thereby reduced stress.Errands were also carefully organised to ensure the shortestwalking distance.

Community walking facilitated by a novel or enjoyablesituation was discussed by a number of people with PDand supported by their partners. Specifically, participantsdescribed reduced symptoms and less fatigue while travellingon holiday than they generally experienced at home, a changewhich could last for a number of weeks after their return.


“Going back three years when (my wife) I’dsay had full blown Parkinson’s, she was very,very bad. We took an overseas trip and . . . (mywife) just kept going and going. By the time wegot to France I flaked. . . She still kept going. . .Something kept her going because as soon as wegot home, boom, she got Parkinson’s again, butwhile we were away it didn’t seem to affect her”(Pa-15).

Optimising pharmaceutical or surgical interventions wasa strong facilitator for some people. Optimal medicationregimes were related to a more efficient gait pattern andless fatigue making long-distance walking more feasible.A positive response to surgical intervention had allowedone participant “freedom” from a schedule of medicationallowing community outings to occur at times convenient forreasons other than medication effectiveness.

“I love it, I love the independence and I lovebeing able to go to the shops and not be dictatedby the medication” (PD-14).

3.1.2. External Facilitators. People with PD and their part-ners reported that partners supported walking in thecommunity by encouraging their partners to go out, bypromoting the importance of continuing to walk as able,by providing physical assistance to overcome barriers in theenvironment, and by supporting the use of attention orcueing strategies. To be effective, cuing strategies needed tobe discrete, mutually agreed on, and practiced to avoid usinga counterproductive cue.

Using equipment was discussed by only a few partici-pants but included changing to more appropriate footwearand carrying a wheelchair in the car in case a longwalking distance or an ineffective dose of medication wasencountered.

Only one aspect of the physical environment wasdescribed as a facilitator to community walking, but thiswas reinforced by many participants. Signalled pedestriancrossings reduce attention required to monitor traffic anddecide when to safely cross and were thereby reportedto facilitate walking in the community. For a number ofparticipants, this had become a habit, now done withoutcompromise.

“. . . you never try to run a light, you always waitfor the lights and you don’t cross any road ifthere is not a light” (Pa-11).

3.2. Barriers. Barriers is the term used to describe factorsreported to exacerbate the negative features of their gait suchas slow walking speed and, therefore, negatively influencethe experience of walking in the community or causeparticipants to avoid walking in the community. Externalenvironmental factors were more frequently perceived tolimit community walking than internal personal factors.

3.2.1. External Barriers. Crowded environments were over-whelmingly disliked by most people in four of the focus

groups. The exception was the rural group in which onlyone participant reported any particular difficulty in crowds.Participants described the need to change direction andavoid obstacles when walking in cluttered (e.g., restaurant)or heavily populated environments (e.g., shopping malls) asa trigger for short shuffling steps and more frequent episodesof freezing. Environments that are busy with people, whoseactions are unpredictable, were the most frequently reportedbarrier.

“I find it more difficult when there are a lot ofpeople around, it means you have to take shortersteps, I like taking long steps, I can balancemyself better” (PD-6).

Attention-demanding environments such as unfamiliarenvironments and road crossing were not reported to con-tribute to any specific gait difficulty, but many participantsreported a need to take extra care while walking in suchenvironments. Road crossing was a particular problem forthe rural group, which was conducted in a town that hadno signalled and very few designated crossings which wereinconveniently located forcing people to cross a busy highwaywithout designated pedestrian crossings.

“Just watching for the traffic—you might not bewalking as quick as you should be and you’rewatching for the traffic. You have to be prettycareful here” (PD-19).

Characteristics of the walking surface such as unevenfootpaths, hills, ramps, flat and inclined moving walkways(travelators), and slippery surfaces were reported as a causeof increased fatigue (hills), fear of falling (uneven andslippery surfaces), and more frequent freezing episodes(ramps and travelators). Even the camber of the footpath,designed to allow water to drain, was commonly reported tomake walking more difficult.

“My greatest difficulty when I’m walking isgoing downhill—can’t handle it, I can go uphillflat out, but I can’t handle going downhill. Evenwith a trolley my feet get stuck on top of a rampand I can’t get going” (PD-2).

The rural group specifically emphasised this barrier. Inthis rural town, footpaths are often absent, where presentsome of the footpaths are tiled and slippery when wet, andthe gutters very deep (20–25 cm high) making access fromthe road to the footpath difficult.

Inclement weather and reduced or fluctuating lightingwere reported to increase difficulty of walking and fear offalling. For some participants these were reasons to avoidcommunity walking all together.

“We avoid going out when it’s raining. It makeshim want to walk faster and he gets so fast thathe shuffles” (Pa-10).

3.2.2. Speed Demands. Only a small number of participantsreported difficulty walking as fast as the environment


demanded. This was often associated with an inability towalk quickly enough to cross the road. One partner reportedthat his wife felt unable to walk quickly enough for him toachieve exercise benefits so she no longer walked with himfor exercise.

“I’m not a quick walker, but it’s quicker than sheis and I don’t mind walking slower but she feelsshe is holding me back . . . that I’m not gettingthe exercise” (Pa-33).

Walking distance was described as a barrier only inthe rural group. Often these participants related greaterwalking distance to greater fatigue and avoided walking inthe community if long distances were encountered.

“because (my partner) can’t walk or stand fora long time, if we can’t get a park close tosomewhere where we want to go we just comehome” (Pa-18).

3.2.3. Internal Barriers. Participants reported that theirresponse to PD medication was unpredictable and walkingwhen medication was not effective very difficult. For someparticipants this meant that trips needed to be postponed,modified, or abandoned due to an ineffective dose.

“I’ll say, right, we’re going down to the shops inhalf an hour—take medication, might get to theshops, medication doesn’t work—(we have to)come home” (Pa-13).

Even with predictable “ON” and “OFF” times, one par-ticipant with PD reported that her need to schedule outingsfor times that medication would be effective gave her a feelingof being “locked to the medication” (PD-14). This on-offphenomenon was also reported as one source of anxiety.

“What if I get weak, what if I can’t move, what ifI’ve got to come home straight away?” (Pa-13).

Anxiety was reported to increase symptoms of PD,resulting in walking difficulty such as shortened step lengthand increased “shuffling” or dragging a leg. Feeling hurried,examined, stigmatised, or judged was also reported toincrease anxiety.

“. . . walking down here this morning I thoughtI would be late and I started dragging my footagain” (PD-19).

Some participants reported fatigue due to longer thanusual walking distance or time. As a result of fatigue, peoplereported abandoning some outings before they had intendedor experiencing fatigue-related weakness and a resultantincrease in walking difficulty.

“You get a fatigue coming in. You will notice itin a weaker muscle group—you might pick it upin the calf where you use it a lot. You might pickit up a hamstring or the front of the leg where itjust becomes harder” (PD-17).

3.3. Disease without Disability. The final theme that emergedis that while strategies and facilitators are effective atovercoming barriers to community walking, people livingwith PD may not appreciate or report any actual problemsor difficulty but rather modifications they have made totheir walking. This suggests that despite the presence ofdisease and impairment some people with PD are able to usefacilitators and strategies to overcome barriers to communitywalking so effectively that no difficulty or disability isconsciously appreciated, even by their partners.

“I find it is not difficult, you just have to becareful in shopping centres with people left rightand centre and you have to keep on the straightand narrow and put your foot in the right place”(PD-2).

“You haven’t had a problem really, have you? Youjust have to think about it” (Pa-22).

It is clear, however, from the barriers outlined abovethat some people with PD are aware of difficulties theyface walking in the community, and some reported verysignificant walking disability.

“I don’t go out on my own (anymore), I have acarer who takes me out” (PD-27).

Which indicates that for some people with PD barriersbecome too significant to overcome using strategies andfacilitators, and disability becomes appreciable.

4. Discussion

Walking has been reported to be the first activity of dailyliving that people with PD identify as having difficultywith, followed closely by a number of activities dependenton walking such as travelling and shopping [15]. To ourknowledge this is the first paper published with such a broadfocus, where the term community walking is used to capturewalking in the community for all reasons including but notexclusive to exercise or physical activity, activities of dailyliving, and leisure activities. Research in other populationshas investigated personal and environmental barriers andfacilitators to physical activity [16].

The results demonstrate that people living with PDappreciate that the ability to walk in the community isthe result of a successful interaction between themselves,including their disease and associated impairments, andthe environment (physical and social) in which they walk.Factors reported to negatively influence this relationshipwere primarily dimensions of the physical environmentwhich previous authors have labelled density (crowding andclutter in the environment), attention, terrain, ambience(weather and lighting), and temporal demands [10]. Notonly do these dimensions present challenges for peoplewith walking impairment, but for people living with PDcertain dimensions can exacerbate the negative featuresof gait. For example, having to stop walking and changedirection while walking in crowded environments demands


frequent stopping, starting, and changing direction, thereby,not allowing people to walk at their preferred speed.This dimension may be particularly challenging for peoplewho experience freezing of gait as turning and negotiatingobstacles are known triggers for freezing [17]. In addition,monitoring the environment for obstacles while walkingmay divert attention away from walking, something thatlaboratory testing has demonstrated people living with PDhave particular difficulty with [5–9].

The results also suggest that the interaction is furthercomplicated for people living with PD whose impairmentsare not static but may fluctuate significantly dependingon the effect of their medication, anxiety, and fatigue. Inone qualitative study of fatigue in people living with PD,all participants agreed that fatigue had a significant anddeleterious effect on their daily activities, social and leisuretime [18]. Two types of fatigue are problematic for peoplewith PD, peripheral and central fatigue [19]. Peripheralfatigue was discussed here as fatigue related to increasedwalking distances, and muscle fatigue related to overuse.Central fatigue is poorly understood and not discussedamong any of these groups. Possibly people who sufferfrom central fatigue are less inclined to commit to outingsand were, therefore, inadvertently excluded from this study.This may also be true of depression, which was also notmentioned in any of these groups.

This sample also reported factors that facilitated walkingin the community. Primarily these facilitators were internalto the person and involved modifying their behaviour orusing strategies to overcome barriers and exploit extrinsicfacilitators so they may continue to walk in the community.For many this behaviour modification is so successful that,despite the presence of disease, disability or difficulty is notperceived. This phase between disease and disability may beconsistent with the phase of preclinical disability experiencedduring aging [20]. In older adults, preclinical disability ischaracterised by reports of no difficulty performing a partic-ular task, but rather reports of modification in the method orfrequency of performing that task [20]. People who reportedhaving modified how or how often they walked half a mile orclimbed ten steps were found to be 3-4 times more likely todevelop disability in the subsequent eighteen months [21].

The current study of walking in community environ-ments adds to a recent qualitative study by Jones et al. [22]which focused on understanding challenges and strategies foreveryday walking in people with PD. Jones et al. asked peoplewith PD to reflect on the challenges and strategies they usedto address the challenges to walking, both indoors and out.Walking whilst doing something else and walking in differentenvironments were two factors identified to increase the chal-lenge of walking. Specifically, participants strongly dislikedwalking in busy and crowded environments. Participants inthat study also described two attention-based strategies thattheir sample described using to improve their walking; thesewere consciously monitoring their walking performance anddirecting attention to correct their gait pattern.

Although some findings are similar, this study differed tothe Jones et al. study in a number of ways. The focus of thisstudy was specifically community walking, and as such the

community-specific barriers and facilitators are presented inmuch greater detail, particularly the environmental barriers.Data was collected using focus groups, rather than in-depthinterviews which may have yielded greater reflection onthe topic, particularly by those participants who reportedmodifications to their walking without appreciable difficultyor disability. The participants of partners in the groups mayhave also contributed to this reflection. Finally this study wasa stand-alone qualitative study with broad inclusion criteria.As such, people with unstable and unpredictable “on” and“off” phenomenon, dyskinesias, and significant walkingdisability were included. Participants were on average 10years after diagnosis with 44% reporting freezing and one-third falling in the past 6 months. The findings reflect theattitudes of those who currently access the community, andas such the results may not be generalisable to all people withPD in all stages of the disease process.

Assessment of community walking has previously beeninferred by assessing an individual’s gait speed and en-durance in an uncluttered environment [11]. The findingsof this study suggest that assessment tasks that incorporatepotentially challenging environmental dimensions such asdensity, attention demands, terrain characteristics, or ambi-ence could provide more specific information about theparticular demands for an individual and how they modifytheir gait to cope. Self-report tools such as the ambulatoryself-confidence questionnaire (ASCQ) [23] and the envi-ronmental analysis of mobility questionnaire (EAMQ) [24]do address some of these issues; however, their accuracyand utility in the PD population is yet to be examined.Furthermore, self-report and actual ability may not alwayscorrelate. Mobility test batteries have been developed toreflect some demands of community mobility [25] but maynot include situations that people with PD find challengingor can include tasks that may be inappropriate.

Wearable sensors such as pedometers, gyroscopes, andaccelerometers have been used to demonstrate changesin activity in people with PD [26–28]. These and othertechnologies have the potential to be developed to measurepeople with PD walking in challenging environments andto possibly monitor their performance when walking in thecommunity.

Therapeutic intervention to manage, prevent, or delaycommunity walking disability is equally complex. The resultsof this study suggest that for people with PD the primarybarriers are external environmental factors. Although advo-cacy for modifying or planning environments that wouldbe more easily negotiated by people with PD may go someway to improve the ability of people with PD to walk in thecommunity; environmental modification may be less feasiblein the community than in a home environment. As such,a more individualised approach to intervention may focuson enhancing likely personal facilitators. This could includeeducating about barriers, facilitators and sharing successfulstrategies used by others, in addition to promoting the useof internal strategies such as attention to walking speed andstep length, and planning for outings. Evidence for the useof interventions to improve community mobility in peoplewith PD is needed.


5. Conclusion

This study reports the perspectives of people with PDand highlights the effectiveness of personal strategies andfacilitators to enable people with PD to continue walking inthe community. People with PD often find environmentalchallenges barriers to walking in the community but do nottend to report disability; rather, they modify their behaviour.Current clinical methods of assessing community mobilitywhich focus on gait speed or distance, thus, may not providesufficient information to accurately reflect a person’s abilityto walk in the community. Furthermore, a deeper under-standing of preclinical walking disability, in people with PD,may allow therapists to provide more timely assessment andtherapy, thereby, delaying the onset of disability rather thanattempting to reverse disability after it presents.

Acknowledgment

The authors would like to acknowledge the contribution ofParkinson’s Queensland Inc. in recruitment of participantsfor this study.

References

[1] G. Alves, E. B. Forsaa, K. F. Pedersen, M. Dreetz Gjerstad, andJ. P. Larsen, “Epidemiology of Parkinson’s disease,” Journal ofNeurology, vol. 255, supplement 5, pp. 18–32, 2008.

[2] O. Blin, A. M. Ferrandez, and G. Serratrice, “Quantitativeanalysis of gait in Parkinson patients: increased variability ofstride length,” Journal of the Neurological Sciences, vol. 98, no.1, pp. 91–97, 1990.

[3] M. E. Morris, R. Iansek, T. A. Matyas, and J. J. Summers, “Thepathogenesis of gait hypokinesia in Parkinson’s disease,” Brain,vol. 117, no. 5, pp. 1169–1181, 1994.

[4] M. Morris, R. Iansek, T. Matyas, and J. Summers, “Abnormali-ties in the stride length-cadence relation in parkinsonian gait,”Movement Disorders, vol. 13, no. 1, pp. 61–69, 1998.

[5] K. Baker, L. Rochester, and A. Nieuwboer, “The immediateeffect of attentional, auditory, and a combined cue strategyon gait during single and dual tasks in Parkinson’s disease,”Archives of Physical Medicine and Rehabilitation, vol. 88, no.12, pp. 1593–1600, 2007.

[6] G. Yogev, N. Giladi, C. Peretz, S. Springer, E. S. Simon, and J.M. Hausdorff, “Dual tasking, gait rhythmicity, and Parkinson’sdisease: which aspects of gait are attention demanding?”European Journal of Neuroscience, vol. 22, no. 5, pp. 1248–1256, 2005.

[7] R. Galletly and S. G. Brauer, “Does the type of concurrenttask affect preferred and cued gait in people with Parkinson’sdisease?” Australian Journal of Physiotherapy, vol. 51, no. 3, pp.175–180, 2005.

[8] J. M. Hausdorff, J. Balash, and N. Giladi, “Effects of cognitivechallenge on gait variability in patients with Parkinson’sdisease,” Journal of Geriatric Psychiatry and Neurology, vol. 16,no. 1, pp. 53–58, 2003.

[9] S. O’Shea, M. E. Morris, and R. Iansek, “Dual task interferenceduring gait in people with Parkinson disease: effects of motorversus cognitive secondary tasks,” Physical Therapy, vol. 82, no.9, pp. 888–897, 2002.

[10] A. E. Patla and A. Shumway-Cook, “Dimensions of mobility:defining the complexity and difficulty associated with commu-nity mobility,” Journal of Aging and Physical Activity, vol. 7, no.1, pp. 7–19, 1999.

[11] S. E. Lord, K. McPherson, H. K. McNaughton, L. Rochester,and M. Weatherall, “Community ambulation after stroke: howimportant and obtainable is it and what measures appearpredictive?” Archives of Physical Medicine and Rehabilitation,vol. 85, no. 2, pp. 234–239, 2004.

[12] C. Peel, P. S. Baker, D. L. Roth, C. J. Brown, E. V. Bodner, and R.M. Allman, “Assessing mobility in older adults: the UAB Studyof Aging Life-Space Assessment,” Physical Therapy, vol. 85, no.10, pp. 1008–1019, 2005.

[13] A. Shumway-Cook, A. Patla, A. Stewart, L. Ferrucci, M.A. Ciol, and J. M. Guralnik, “Environmental componentsof mobility disability in community-living older persons,”Journal of the American Geriatrics Society, vol. 51, no. 3, pp.393–398, 2003.

[14] J. Green and N. Thorogood, Qualitative Methods for HealthResearch, London, UK, 2nd edition, 2009.

[15] L. M. Shulman, A. L. Gruber-Baldini, K. E. Anderson et al.,“The evolution of disability in Parkinson disease,” MovementDisorders, vol. 23, no. 6, pp. 790–796, 2008.

[16] M. A. Plow, L. Resnik, and S. M. Allen, “Exploring physicalactivity behaviour of persons with multiple sclerosis: a quali-tative pilot study,” Disability and Rehabilitation, vol. 31, no. 20,pp. 1652–1665, 2009.

[17] B. R. Bloem, J. M. Hausdorff, J. E. Visser, and N. Giladi, “Fallsand freezing of gait in Parkinson’s disease: a review of twointerconnected, episodic phenomena,” Movement Disorders,vol. 19, no. 8, pp. 871–884, 2004.

[18] R. G. Brown, A. Dittner, L. Findley, and S. C. Wessely, “TheParkinson fatigue scale,” Parkinsonism and Related Disorders,vol. 11, no. 1, pp. 49–55, 2005.

[19] I. Ziv, M. Avraham, Y. Michaelov et al., “Enhanced fatigueduring motor performance in patients with Parkinson’sdisease,” Neurology, vol. 51, no. 6, pp. 1583–1586, 1998.

[20] L. P. Fried, K. Bandeen-Roche, P. H. M. Chaves, and B.A. Johnson, “Preclinical mobility disability predicts incidentmobility disability in older women,” Journals of Gerontology,vol. 55, no. 1, pp. M43–M52, 2000.

[21] L. P. Fried, Y. Young, G. Rubin, and K. Bandeen-Roche, “Self-reported preclinical disability identifies older women withearly declines in performance and early disease,” Journal ofClinical Epidemiology, vol. 54, no. 9, pp. 889–901, 2001.

[22] D. Jones, L. Rochester, A. Birleson et al., “Everyday walkingwith Parkinson’s disease: understanding personal challengesand strategies,” Disability and Rehabilitation, vol. 30, no. 16,pp. 1213–1221, 2008.

[23] M. Asano, W. C. Miller, and J. J. Eng, “Development andpsychometric properties of the ambulatory self-confidencequestionnaire,” Gerontology, vol. 53, no. 6, pp. 373–381, 2007.

[24] A. Shumway-Cook, A. Patla, A. L. Stewart, L. Ferrucci, M.A. Ciol, and J. M. Guralnik, “Assessing environmentallydetermined mobility disability: self-report versus observedcommunity mobility,” Journal of the American GeriatricsSociety, vol. 53, no. 4, pp. 700–704, 2005.

[25] J. A. Howe, E. L. Inness, A. Venturini, J. I. Williams, and M.C. Verrier, “The Community Balance and Mobility Scale—abalance measure for individuals with traumatic brain injury,”Clinical Rehabilitation, vol. 20, no. 10, pp. 885–895, 2006.


[26] L. A. Hale, J. Pal, and I. Becker, “Measuring free-living physicalactivity in adults with and without neurologic dysfunctionwith a triaxial accelerometer,” Archives of Physical Medicine andRehabilitation, vol. 89, no. 9, pp. 1765–1771, 2008.

[27] S. F. Chastin, K. Baker, D. Jones, D. Burn, M. H. Granat, and L.Rochester, “The pattern of habitual sedentary behavior is dif-ferent in advanced Parkinson’s disease,” Movement Disorders,vol. 25, no. 13, pp. 2114–2120, 2010.

[28] B. Dijkstra, Y. P. Kamsma, and W. Zijlstra, “Detection ofgait and postures using a miniaturized triaxial accelerometer-based system: accuracy in patients with mild to moderateParkinson’s disease,” Archives of Physical Medicine and Reha-bilitation, vol. 91, no. 8, pp. 1272–1277, 2010.


Review Article

Progressive Resistance Exercise and Parkinson’s Disease:A Review of Potential Mechanisms

Fabian J. David,1 Miriam R. Rafferty,2 Julie A. Robichaud,1 Janey Prodoehl,1

Wendy M. Kohrt,3 David E. Vaillancourt,4, 5 and Daniel M. Corcos1, 6, 7, 8

1 Department of Kinesiology and Nutrition, University of Illinois at Chicago, Chicago, IL 60612, USA2 Graduate Program in Neuroscience, University of Illinois at Chicago, Chicago, IL 60612, USA3 Division of Geriatric Medicine, University of Colorado School of Medicine, Aurora, CO 80045, USA4 Department of Applied Physiology and Kinesiology, University of Florida, Gainesville, FL 32611, USA5 Department of Neurology, University of Florida, Gainesville, FL 32610, USA6 Department of Bioengineering, University of Illinois at Chicago, Chicago, IL 60612, USA7 Department of Physical Therapy, University of Illinois at Chicago, Chicago, IL 60612, USA8 Department of Neurological Sciences, Rush University, Chicago, IL 60612, USA

Correspondence should be addressed to Fabian J. David, [email protected]

Received 20 July 2011; Revised 19 September 2011; Accepted 20 September 2011


Copyright © 2012 Fabian J. David et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This paper reviews the therapeutically beneficial effects of progressive resistance exercise (PRE) on Parkinson’s disease (PD). First,this paper discusses the rationale for PRE in PD. Within the first section, the review discusses the central mechanisms that underliebradykinesia and muscle weakness, highlights findings related to the central changes that accompany PRE in healthy individuals,and extends these findings to individuals with PD. It then illustrates the hypothesized positive effects of PRE on nigro-striatal-thalamo-cortical activation and connectivity. Second, it reviews recent findings of the use of PRE in individuals with PD. Finally,knowledge gaps of using PRE on individuals with PD are discussed along with suggestions for future research.

1. Introduction

The standard treatment for Parkinson’s disease (PD) is phar-macologic treatment with levodopa, a precursor to do-pamine. However, continued treatment with levodopa isassociated with motor side effects such as dyskinesias andmotor fluctuations. Until an oral formulation of levodopawithout the accompanying motor side effects is formulated,surgical options offer some relief. Typically, surgery isreserved for when the disease and the side effects dueto medication are severely disabling. Currently, the mostcommon surgical option is high-frequency deep brainstimulation of the subthalamic nucleus or the internal globuspallidus [1–4]. Despite the substantial clinical benefits ofsurgery, surgical treatment is not without complications,which occur in up to 50% of individuals with PD whoundergo deep brain stimulation [2, 5]. These complicationsinclude device/surgery-related infections, cognitive decline,

depression, speech difficulties, gait disorders, and posturalinstability [2, 5]. Therefore, there is merit to exploring treat-ment options that may be used as adjuncts to pharmacologicand surgical treatments prescribed in PD. One such option isexercise, specifically progressive resistance exercise (PRE).

This review paper will first discuss the rationale for PREin PD specifically related to bradykinesia and muscle weak-ness. Then it will review recent findings related to the use ofPRE in individuals with PD. Finally, it will identify gaps inknowledge of using PRE in individuals with PD and makessuggestions for future research.

2. Rationale for ProgressiveResistance Exercise

This section will set up the basis for PRE as a therapeuticintervention in PD. To do so, we will outline the underlyingmechanisms for the motor symptoms that can be treated


with PRE. We will focus primarily on the central mechanismsthat underlie bradykinesia and muscle weakness in PD.Then we will discuss the central changes that accompanyPRE and hypothesize how these changes might modify thecentral mechanisms that underlie bradykinesia and muscleweakness. We will conclude this section with our rationalefor the use of PRE in individuals with PD.

2.1. Bradykinesia and Muscle Weakness. Bradykinesia refersto the slowness of a performed movement [6]. Bradykinesiais a primary motor symptom of PD, which is also consid-ered the most functionally debilitating symptom and is aconsistent feature of the disease [7]. Muscle weakness, whichis a reduction in the amount of force generated by musclecontraction, is often observed in individuals with PD. In fact,several studies have demonstrated that individuals with PDexhibit muscle weakness [8–15]. We have shown that thisweakness is exaggerated in the extensor muscles, specificallyextensors of the elbow [8, 16]. Additionally, muscle weaknesshas also been observed across various muscle groups in thetrunk [11], upper limbs [14], and lower limbs [9, 10, 13, 14].

In PD, the idea that bradykinesia and weakness arerelated can be derived from the fact that bradykinesiaand muscle weakness might share common underlyingmechanisms. Central to the pathophysiology of PD is theknown nigral dopaminergic deficit that results in an increasein tonic inhibition of the thalamus and reduction in theexcitatory drive to the motor cortex [17]. This, in turn, mayresult in disruption of the cortical activation of the muscle[18–21] and may manifest as bradykinesia and muscleweakness. Further, muscle power, the product of movementvelocity and muscle torque, is reduced in individuals withPD [13]. Also, torque production during isokinetic musclestrength testing in individuals with PD has been shown tovary with movement velocity. Nogaki et al. found that inindividuals with PD, no difference was observed in peaktorque between the more and the less affected side for slowermovements, while for faster movements, the more affectedside was significantly weaker than the less affected side [22].Therefore, reduction in muscle power is indicative of deficitsin either strength, movement speed, or both and strengthensthe proposed relationship between bradykinesia and muscleweakness.

Given that the muscle is the final target of corticaloutput during movement and force production, analyzingthe electromyographic (EMG) activation patterns can pro-vide insight into hypothesized impairments that underliebradykinesia and muscle weakness. We have shown thatin individuals with PD, EMG activation patterns duringballistic movements and isometric actions are abnormal andreflect impaired activation of the muscle. Muscle activationpatterns during ballistic movement in individuals with PDare abnormal in four significant ways. First, muscle activa-tion patterns show increased variability when compared toage- and sex-matched healthy individuals [23, 24]. Second,in contrast to healthy individuals, the first agonist burstduration does not systematically increase with movementdistance [23]. Third, the magnitude of the first agonistburst, early in the disease, is similar to that observed in

healthy individuals; however, as the diseases progresses,the magnitude of the first agonist burst is modulated lesswith increasing movement distance [23]. Fourth, multipleagonist bursting is observed during the acceleration phase ofmovement, and the number of agonist bursts increases withincreasing the movement distance [23, 24]. During isometricactions, individuals with PD manifest deficits throughoutthe task. At the very beginning of the task, they exhibitdecreased rate of torque generation and decreased initialphasic agonist EMG activation, which results in prolongedtorque rise times and delayed peak torque [16]. In the middleof the task, during steady-state contraction at 25%, 50%,and 75% of maximal voluntary contraction (MVC), thedominant frequency in the EMG spectrogram in individualswith PD stays fairly constant at ∼10 Hz [25]. In healthyindividuals, however, the dominant frequency is higher andincreases with the increase in isometric torque generation,that is, the dominant frequency shifts from ∼18 to 25 Hzwhen isometric torque generation increases from 25% to75% of MVC [25]. At the end of the task, the rate of releaseof muscle contraction is also prolonged, and torque fall timesare increased in individuals with PD [26].

The abnormal EMG activation patterns discussed abovecan be partly explained in terms of an impairment in thecorticospinal activation of the muscle, specifically, impair-ments in variability, intensity, and frequency of the corti-cospinal activation of the muscle. Increased variability inthe corticospinal activation of the muscle could lead tovariability in motor unit recruitment and result in increasedEMG variability [27]. This increased variability in motorunit recruitment could impair coordinated relaxation ofactively contracting motor units, contributing to prolongeddeceleration phases during movement and prolonged relax-ation times during isometric torque generation. Reductionin the intensity of the corticospinal activation of the muscle[28] may result in impaired motor unit recruitment andcould contribute both to bradykinesia and muscle weakness.For instance, impaired motor unit recruitment duringmovement could result in reduced angular impulse duringthe acceleration phase of a movement and contribute tobradykinesia, and impaired motor unit recruitment duringisometric torque generation could result in reduced peaktorque and contribute to muscle weakness.

Alterations in the frequency of the corticospinal activa-tion of the muscle could also explain some of the abnor-mal EMG patterns observed in individuals with PD. Inhealthy subjects the corticospinal activation to the muscleis characterized by three primary frequencies, that is, 10 Hz,20 Hz, and 40 Hz [29, 30]. The magnetoencephalic (MEG)power spectrum is dominated by ∼20 Hz oscillations duringweak contractions and ∼40 Hz oscillation during strongcontractions [29]. Similarly, the mean power in the EMGpower spectrum increases from 10 Hz to 25 Hz with increasein percent MVC from 10% to 80% of MVC [31]. Inuntreated (de novo) individuals with PD relative to age-and sex-matched healthy individuals, resting state corticalactivity in the 8–10 Hz band is increased, while activity inthe 30–48 Hz band is reduced [32]. Further, in individualswith PD, the EMG power spectrum is dominated by power


in the low-frequency band (∼10–15 Hz) [25, 26, 29], and theMEG-EMG coherence is strong in this low-frequency bandwith the MEG signal leading the EMG signal by ∼15–38 ms[29]. Thus, one could hypothesize that if the cortical signalto the muscle is dominated by low-frequency oscillations,then this limits the ability to recruit larger, high-frequencymotor units, which are required to rapidly generate torqueduring ballistic movements and generate maximal torqueduring isometric torque generation. The evidence reviewedin this and the previous two paragraphs suggests that EMGpatterns are abnormal in individuals with PD, and one likelyexplanation for these observed EMG abnormalities is deficitsin the variability, intensity, and frequency of the corticospinalactivation of the muscle.

Another factor that could contribute to muscle weaknessin individuals with PD is reduced muscle mass. Evidencethat muscle mass is reduced in PD is provided by Petroniand colleagues [33]. They reported that midarm musclecircumference was below the 10th percentile in 23% ofindividuals with advanced PD between 65 and 75 years ofage [33]. On the other hand, evidence that this is not thecase is provided by Markus and colleagues [34]. They foundthat even though body mass index and skin fold thickness,relative to age- and sex-matched healthy individuals, werereduced in individuals with PD, midarm circumference wasnot different from healthy individuals. Thus, the authorsconcluded that decrease in body mass index was due to a lossof fat and not due to a loss of muscle mass.

It is important to note that not only does PD causeweakness, but it is highly likely that muscle weakness andfunctional limitations such as postural instability and gaitdisturbances lead to reduced physical inactivity as a com-pensatory mechanism to minimize the likelihood of falls[35]. Therefore, physical inactivity can contribute to muscleweakness and lead to a vicious cycle between muscleweakness and physical inactivity [36].

Even though we cannot discount muscle mass andchanges in muscle properties as likely contributors to muscleweakness, it is our stand that the primary contributors tomuscle weakness are central in origin and are related todopaminergic deficits. This is evidenced by the fact that bothanti-Parkinsonian medication and deep brain stimulationresult in significant improvement in movement speed [24,37] and significant gains in muscle strength in relatively shortamounts of time (not longer than 90 minutes) [16, 38, 39].Given that the minimum amount of time required to noticeappreciable hypertrophy is at least 20 days [40], it is highlyunlikely that the immediate strength gains brought about byanti-Parkinsonian medication or deep brain stimulation arecaused by gains in muscle mass.

The question that remains is the extent to which bradyki-nesia and weakness can be compensated for. We have shownthat levodopa and/or deep brain stimulation of the subthala-mic nucleus improves bradykinesia and/or muscle strength[24, 38, 39]; however, bradykinesia is not normalized [24,37]. Moreover, surgical interventions carry significant risks,while medication becomes progressively less effective, andthe side effects of medication get progressively worse overtime. Therefore, until a cure for PD can be identified, there

is a compelling need to develop interventions that improvethe signs and symptoms of the disease and slow down therate at which the signs and symptoms of the disease worsen.One such intervention is PRE, which may be a beneficialand cost effective adjunct treatment in managing PD. Assuch, if PRE is to be beneficial for individuals with PD,it should bring about central changes that potentially alternigro-striatal-thalamo-cortical activation and connectivity.Since this has not yet been studied in individuals with PD,we will discuss the central changes that accompany PREin healthy young and elderly individuals and extend thesefindings to individuals with PD.

2.2. Central Changes That Accompany Progressive ResistanceExercise. The evidence for the central changes that accom-pany PRE is threefold [41]. First, gains in muscular strengthappear before noticeable muscle hypertrophy [41, 42]. Aftercommencing a PRE protocol, strength gains appear asearly as 5 days [43], but muscle hypertrophy appears noearlier than 20 days [40]. Therefore, the initial gains inmuscle strength cannot be explained by measurable musclehypertrophy. Instead, a likely explanation for the observedstrength gains is the central changes that accompany PRE.Second, cross-education (i.e., improved performance in theuntrained limb) is often observed [41]. Munn and colleagues,in their meta-analysis that included 13 studies, concludedthat unilateral PRE brings about a 7% increase in strengthin the untrained contralateral limb [44]. Given that thiscross-education effect is accompanied by increase in musclesurface EMG, but is not accompanied by gains in musclesize, it is likely to be brought about by the central changesthat accompany PRE [42, 45]. Third, improvements in per-formance following PRE are both specific and generalized.The argument for specificity arises from the fact that short-term dynamic strength training results in significantly greatergains in dynamic strength, while isometric strength gains aremarginal [46]. While the argument for generalizability arisesfrom the fact that short-term strength training that focuseson increasing isometric strength also improves movementcoordination during an untrained task [47]. Thus, bothspecific and generalizable motor learning effects of PREprovide a third line of evidence for the central changes thataccompany PRE.

Further evidence for the central changes that accompanyPRE comes from studies employing transcranial magneticstimulation (TMS), electroencephalography (EEG), func-tional magnetic resonance imaging (fMRI), and muscle EMGactivation patterns. Using TMS, Carroll and colleagues foundthat for the same level of torque, the amplitude of the motorevoked potential was significantly reduced following a 4-week PRE program [48]. They concluded that resistancetraining altered the functional properties of the spinal cordcircuitry, and fewer motor neurons were recruited for similarlevels of pretraining torque. Using EEG, Falvo and colleaguesfound that the movement-related cortical potentials weresignificantly attenuated following a 3-week PRE program[49]. They concluded that PRE reduced the neural effortrequired to move similar levels of pretraining loads. UsingfMRI, Liu-Ambrose and colleagues found that in elderly


women, following PRE, percent signal change significantlyincreased in the left anterior insula and the anterior portionof the left middle temporal gyrus [50]. They concluded thatPRE could facilitate functional plasticity in the cortex. UsingEMG, several studies have shown that muscle activationpatterns change after PRE [42, 49, 51–54]. These muscleactivation changes following PRE include an increase inthe EMG activation [40, 53, 54], possibly due to increasedmotor unit recruitment [55–57], increased firing rate [57,58], and improved synchronization [52, 59]; a reduction inthe EMG activation to torque ratio, that is, reduction inEMG activation relative to the amount of torque produced[60]; a reduction in the variability associated with thetiming, amplitude, and duration of muscle activity [47]; areduced agonist-antagonist coactivation [61]. In addition,central changes accompanying PRE have been inferred usingthe H-reflex to examine motor neuron reflex excitability.Holtermann and colleagues found that the amplitude ofthe H-reflex increased following a 3-week PRE program inhealthy individuals [62]. Further, they found that the H-reflex increase in amplitude was associated with an increasedrate of force development. This could provide a neurophys-iological basis for PRE improving bradykinesia in PD. Theexact mechanisms underlying the observed increase of theH-reflex amplitude are not yet known however. The authorssuggested that one possibility is that the reflex excitability ofthe motor neuron pool may be enhanced following PRE.

It should be noted that some of the neural changesdiscussed in the preceding paragraphs may be affected byfactors such as age, sex, the muscle group trained, and theirinteractions [63, 64]. For instance, following PRE, upperand lower body strength gains are greater in young than inhealthy elderly individuals [63]. Also, upper body strengthgains are greater in men than in women; however, lower bodystrength gains are not different between men and women[63].

In summary, PRE can bring about changes throughoutthe neural axis. Currently, none of the central changesthat accompany PRE discussed previously in this sec-tion have been researched in individuals with PD. Eventhough improvements in neuromuscular function have beenobserved in individuals with PD, from a physiological per-spective, further research is required to elucidate the centralchanges that accompany PRE that could mitigate the motorand nonmotor symptoms observed in PD.

Brain regions where PRE could potentially alter activityinclude the motor cortex, the posterior putamen, the internalglobus pallidus (GPi), and the subthalamic nucleus (STN)(Figure 1). Fisher and colleagues recently demonstratedmotor cortical changes following body-weight-supportedtreadmill training in individuals with PD [65]. Theyshowed that cortical hyperexcitability, which is consistentlyobserved in individuals with PD, is reversed following body-weight-supported treadmill training [65]. Petzinger andcolleagues have also shown an increase in the stimulus-evoked dopamine release within the dorsolateral striatumfollowing intensive treadmill training in 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine- (MPTP-) lesioned mice [66].Because the dorsolateral striatum is engaged to a high degree

during fore- and hind-limb movements during treadmillexercise, they attributed the observed striatal plasticity touse-dependent synaptic plasticity.

Similarly, there may also be use-dependent synapticplasticity in the putamen, the GPi, and the STN followingPRE. Our lab has conducted a series of studies in which wehave shown that nuclei within the basal ganglia scale withthe performance of different force producing tasks in bothhealthy individuals and individuals with PD. Specifically,we have shown that both the globus pallidus and the STNincrease percent signal change when generating progressivelylarger forces in healthy individuals [67]. We have also shownthat individuals with PD have a reduced percent signalchange in all nuclei of the basal ganglia during an isometricforce production task, even early in the disease processwhen individuals have not yet started their anti-Parkinsonianmedication [68]. In addition, blood-oxygen-level-dependentactivity in the nuclei of the basal ganglia was correlated to themotor section of the Unified Parkinson’s Disease Rating Scale(UPDRS) [69]. The symptom with the highest correlationwith basal ganglia activity was bradykinesia. Thus, if PREswere shown to alter the motor section of the UPDRS andbradykinesia, then it is possible that the neuronal activity ofthe basal ganglia would also be altered by PRE.

Figure 1 illustrates the hypothesized positive effects PREmight have in individuals with PD by possibly alteringactivity and connectivity in cortical and subcortical regions.It should be noted that these effects of PRE on activity andconnectivity in cortical and subcortical regions are purelyspeculative, as there are no in vivo studies that have examinedthis relationship. As can be clearly seen from the figurehowever, the basal ganglia are strategically positioned toinfluence cortical output and modulate control of movementand force. As such, we suggest that one potential reason forwhy PRE could be therapeutically beneficial for individualswith PD is that it may alter activity in the cortex and thebasal ganglia, and connectivity between and within theseregions. Advances in experimental techniques, such as TMS,EEG, fMRI, positron emission tomography (PET), diffusiontensor imaging (DTI), and EMG and reflex analyses, affordthe possibility of testing hypotheses related to the effect ofPRE on neural activity, neural connectivity, and structuralintegrity in vivo, in humans. Figure 1 shows the outcomesand tools that can be used to empirically determine theeffects of PRE in specific brain regions. To elaborate, changesin cortical excitability can be measured using TMS, whilechanges in cortical activity and intracortical connectivity canbe measured using EEG. Functional MRI can be used toidentify blood-oxygenation-level-dependant signal changesin cortical and subcortical regions following PRE. PET can beused to investigate the effect of PRE on dopamine synthesis,transport, and usage. Diffusion tensor imaging can helpelucidate hypotheses related to the changes in structure incortical and subcortical regions, namely, the substantia nigra,the STN, and the thalamus. Reflex and EMG analyses canbe used to identify reflex changes, such as change in H-reflex amplitude, and changes in EMG activation patternsto infer central changes following PRE. Prior to embarkingon empirical verification of some of the ideas presented in


Corticospinal excitabilityCortical activityIntracortical connectivityBOLD signal

Dopamine usageBOLD signalStructural integrity

Dopamine transport

Dopamine synthesisBOLD signalStructural integrity

BOLD signalStructural integrity

Outcomes

TMSEEGEEGfMRI

Tools

PET

fMRI

DTI

PET

PETfMRIDTI

fMRIDTI

Reflex modulationMuscle activity

EMGEMG

ExcitatoryInhibitory

SNc

STN

ThalamusCaudate

GPe

GPi

Cortex

Spinal cord

Muscle

Putamen

Figure 1: Hypothesized central effects PRE might have in the cortex, basal ganglia, and spinal cord and the tools that can be used to examinethese hypothesized changes. TMS, transcranial magnetic stimulation; EEG, electroencephalography; fMRI, functional magnetic resonanceimaging; PET: positron emission tomography; DTI: diffusion tensor imaging; EMG: electromyography; SNc: substantia nigra pars compacta;GPe: external globus pallidus; GPi: internal globus pallidus; STN: subthalamic nucleus.

this paragraph, researchers are cautioned on the technicaldifficulties, limitations, and the complications of the above-mentioned methods (for a recent detailed review, see Carrollet al. [70]).

In conclusion, the rationale for the use of PRE in PDis fourfold. First, as discussed above, individuals with PDexhibit muscle weakness. PRE can significantly increase thetorque- and power-generating capacity of the muscle, thusdirectly affecting muscle weakness. Even though other formsof exercise such as aerobic exercise provide substantial healthbenefits, they do not improve muscle strength by design.Improvements in muscle strength and power have significantimpact on bradykinesia [71] and could also facilitate inde-pendence in the community, improve functional mobility,and may reduce the risk of falls [72]. Second, exerciseinterventions in general have been shown to enhance corticalactivity, possibly beneficially altering variability, intensity,and frequency components of the corticospinal activationof the muscle [47–49, 73]. This could significantly impactbradykinesia in individuals with PD [65]. Third, exercise mayslow down the rate at which the UPDRS scores increase. TheUPDRS is the clinical gold standard for assessing the severityand progression of symptoms in PD and for evaluatingnovel therapies, with higher scores reflecting more severedisease. Reuter and colleagues have shown that a 14-week,intense, multimodal exercise training program can bringabout ∼12 point reduction in the motor UPDRS scores[74]. Additionally, physical activity has been associated withincreasing the survival rate of individuals with PD [75].

Finally, there may well be additional benefits for the non-motor symptoms of PD, such as executive function, mood,and quality of life.

3. Progressive Resistance Exercise in PD

Rehabilitation research studies in individuals with PD dem-onstrate that PRE can have a positive effect on muscle size[76], muscle strength [15, 71, 76–78], muscular endurance[77, 79], and neuromuscular function [71, 76–79]. To date,only one study [76] has quantified changes in muscle sizein individuals with PD. Dibble and colleagues observed a6% increase in muscle volume, measured using volumetricmagnetic resonance imaging, after a 12-week eccentric PREprogram [76]. Eccentric PRE training involves the use ofeccentric muscle activity, that is, the active lengthening ofmuscles when an external load is imposed; consequently,work is done on the muscle [80]. The rationale used byDibble and colleagues for using eccentric PRE is that for thesame amount of work (i.e., force × distance), high levels offorce are generated with minimal oxygen consumption [81].

With regard to muscle strength, several studies havedemonstrated significant gains in muscle strength followingPRE in PD [15, 71, 76–78]. For instance, improvementsin strength were observed by Hirsch and colleagues ina randomized controlled trial that compared a 10-weekbalance training protocol to a 10-week balance training plusPRE protocol [78]. At the end of 10 weeks, they observedsignificant improvements in strength in knee extension, knee


flexion, and ankle plantar flexion in the balance plus PREgroup. When the strength measures were combined acrossthe knee and ankle, they observed a 52% increase in strengthfrom before to after treatment in the balance plus PRE group.In another randomized placebo-controlled trial, Hass andcolleagues demonstrated significant gains in strength andendurance in upper body muscles, following a 12-week PREprogram supplemented with creatine monohydrate [77].Improvement in endurance was observed by Scandalis andcolleagues following an 8-week PRE program that was gearedtoward the lower body [79]. They found improvementsin the total number of abdominal crunches that could beperformed at one time. They also observed improvements inlower limb performance, which was quantified as a productof repetitions and weight. Next, we will review the evidencethat supports positive changes in neuromuscular functionthat accompany strength gains in individuals with PD fol-lowing PRE.

From a rehabilitation perspective, it is critical thatstrength gains bring about corresponding improvements inneuromuscular function, such as gait, stair climbing, timedup and go, and postural stability. To this end, recent studieshave shown significant improvement in neuromuscularfunction following PRE interventions in PD. First, improve-ments in gait have been reported. Three-dimensional gaitanalyses following an 8-week PRE program demonstratedthat individuals with PD increased their gait velocity, stridelength, and head angle relative to the floor during midstride[79]. Similar findings of increased gait velocity were alsoreported by Dibble and colleagues following a 12-weekeccentric PRE intervention [71, 76]. The functional gaitoutcomes included the six-minute walk, ten-meter walk,timed up and go, and stair ascent and descent times. Theyobserved that individuals with PD significantly improvedgait velocity and increased the distance walked in six-minutes, reduced the time taken to walk ten meters, reducedthe time taken to complete the timed up and go, and reducedstair descent times. Their findings led them to conclude thatprogressive resistance eccentric exercise could significantlyimpact bradykinesia. Second, improvement in postural sta-bility has been reported. Hirsch and colleagues showed thatindividuals with PD demonstrated an improved ability tomaintain balance during destabilizing conditions following a10-week balance plus PRE intervention [78]. Third, improve-ment in patient-perceived quality of life has been reported.Even though quality of life is not a direct measure of neuro-muscular function, it is reasonable to assume that improvedneuromuscular function might contribute to improved qual-ity of life. Dibble and colleagues found that eccentric PREsignificantly improved patient-perceived quality of life asmeasured by the Parkinson’s disease questionnaire (PDQ-39)[71].

In summary, PRE can significantly improve muscle size,muscle strength, muscle endurance, and neuromuscularfunction and can significantly impact areas often reportedto be problematic in individuals with PD, such as bradyki-nesia, postural instability, and patient-perceived quality oflife.

4. Limitations of Current Research andRecommendations for Future Research

The few studies that have examined the effect of PRE inPD are no doubt vital to our continued understanding ofthe effect of PRE and the pursuit of adjunct treatments forPD; however, they are not without limitations. First, it is notclear how anti-Parkinsonian medications interact with PRE.To ascertain the unique contribution of PRE on strengthand functional outcomes in PD, it is essential to examineindividuals while off anti-Parkinsonian medications. Also,if changes to the underlying disease process are to beevaluated, this is best done while off medication. Among thestudies reviewed, all except for Scandalis and colleagues [79]tested individuals with PD while on medication. Thus, moreresearch is required to investigate the unique effect of PREon outcomes of strength, neuromuscular function, and theunderlying disease process.

Second, the motor UPDRS, which is the clinical goldstandard of assessing severity of motor deficits in PD, hasrarely been used as an outcome measure while evaluating theeffects of PRE. In order to convince neurologists who manageindividuals with PD to prescribe exercise as an adjuncttherapy, it is vital to demonstrate clinically important changeon the motor UPDRS as a result of PRE. Minimal clinicallyimportant change on the motor UPDRS is based on the effectof anti-Parkinsonian medication and is defined as a 5-pointreduction on the motor UPDRS score [82]. The scores onthe motor UPDRS range from 0 to 108, and higher scoresindicate more severe motor symptoms. Thus, if exercise canbring about at least a 5-point reduction in the motor UPDRS,one can make a compelling case to include PRE as an adjunctto the standard management of PD. Future research shouldinclude the motor UPDRS as an outcome measure whileevaluating the effects of PRE. To date, Dibble et al. [71] andHass et al. [77] have used the motor UPDRS as an outcomemeasure; however, they both failed to show any clinicallyrelevant change following PRE. This could have been due tothe fact that these studies tested individuals with PD whileon medication and/or due to the short duration of the PREintervention.

Third, long-term effects of PRE are yet to be determined.All of the studies conducted to date evaluate the effect ofPRE over 8 to 24 weeks. Given that PD is a progressiveneurodegenerative disorder and is further affected by theprocess of aging, which is accompanied by decline in strengthand neuromuscular function [83], it is vital that the long-term effects of PRE are thoroughly understood. For instance,continued benefit of PRE over the long-term could reducethe rate at which the disease progresses. This is significant,especially because recent exciting epidemiological researchhas concluded that moderate to vigorous levels of physicalactivity in mid- or later life may be associated with a 40%reduction in the future risk of being diagnosed with PD [84].Additionally, PRE over the long term could reduce the rate atwhich dosage of medication is increased and possibly delaythe onset of dyskinesias, as well as surgical interventions.Thus, it is essential that future studies evaluate the effects ofPRE over the long term in PD.


Fourth, even though it is accepted that cognitive impair-ment is frequently observed in PD [85–90], the effect ofPRE on cognitive function in PD is not well researched.The rationale for PRE as a therapeutic intervention forcognitive dysfunction is threefold. First, PRE has been foundto improve cognitive function in healthy subjects betweenthe age of 65 and 75. Cassilhas et al. demonstrated improvedperformance on measures of working memory and attentionfor those assigned to 24 weeks of PRE [91]. More recently,Liu-Ambrose and colleagues demonstrated beneficial cog-nitive effects of 52 weeks of PRE in community dwellingelderly women [92]. They showed improvements in attentionand conflict resolution. Additionally, in a subsequent studywith the same sample, they demonstrated changes in percentsignal change in brain areas that correspond to conflict reso-lution [50]. Second, even though aerobic training providescognitive benefits, a combination of aerobic and PRE hasbeen evidenced to render the greatest cognitive benefits [93].Recently, two studies have evaluated the combined effect ofPRE and aerobic exercise on executive function in PD [94,95]. Both studies concluded that PRE combined with aerobicexercise improved executive function. Third, there is a strongbiological basis for the cognitive benefits gained from PRE.These include the reduction in serum levels of homocysteine[96] and the increase in serum levels of insulin-like growthfactor I [97], following PRE, which are both known to beassociated with cognitive function [98, 99]. Thus, there isevidence in the literature to support the beneficial effectsof PRE on cognitive function, and future research shouldaddress this in individuals with PD.

Fifth, the diverse experimental designs employed in thestudies reviewed may be less than ideal. Given the realities ofconducting research with a patient population, the studiesreviewed provide an excellent basis for large-scale, long-term prospective randomized clinical trials. However, thesmall sample sizes used (between 6 and 14 per group,with a total sample size not exceeding 20), the lack ofrater blinding (only Hass et al.’s was a randomized, double-blinded, placebo-controled trial [77]; while Hirsch et al.’s wasa randomized control trial, the raters were unblinded [78]),and not employing the intent-to-treat principle in statisticalanalysis lead to biases that could question the validity ofsome of the conclusions. Thus, future studies should beblinded, randomized clinical trials, which will provide themost robust experimental design to address the gaps in theliterature by assessing the short- and long-term effects of PREin individuals with PD.

Sixth, the optimal PRE prescription for individuals withPD is yet to be established. There are two aspects of treatmentoptimization. The first aspect is the optimization of PREparameters, such as the frequency, intensity, duration, andmode of exercise (i.e., strength and power training). Thesecond aspect is the optimization of PRE with regards to thevarious clinical subtypes of PD. Within the general diagnosisof PD, distinct clinical subtypes have been identified basedin part on the age of onset, the predominant motor sign(e.g., tremor dominant, nontremor-dominant akinetic rigidetc.), and the clinical course of the disease [100]. There isevidence in the literature that suggests that these different

PD subtypes may respond differently to interventions andmay progress at different rates [101–103]. For example,individuals who begin with significant rest tremor may notrespond as well to levodopa and may progress at a slowerrate compared to individuals who present with a nontremor-dominant, akinetic-rigid form of the disease. It is likely thatthe effect of PRE may vary with the clinical subtype of PD.In addition, the effect of PRE on tremor and rigidity isnot yet known. Thus, future research should identify theoptimal PRE prescription in the context of the differentclinical subtypes of individuals with PD and empiricallyverify hypotheses related to tremor and rigidity as well.

5. Conclusion

In PD, bradykinesia and muscle weakness are primarilydue to nigral dopaminergic deficits that alter corticospinalactivation. Given the wide array of neural changes thataccompany PRE summarized in this paper, the potential toslow the rate of the progression of the symptoms of PD,the improvement in strength and function, and the positiveeffects on nonmotor symptoms of PD, there is a strongrationale for the use of PRE as an adjunct treatment in PD.

Acknowledgment

This publication was made possible by the National Institutesof Health (5R01NS028127-16, 5T32MH067631-07).

References

[1] E. Moro, A. M. Lozano, P. Pollak et al., “Long-term results ofa multicenter study on subthalamic and pallidal stimulationin Parkinson’s disease,” Movement Disorders, vol. 25, no. 5,pp. 578–586, 2010.

[2] K. A. Follett, F. M. Weaver, M. Stern et al., “Pallidal versussubthalamic deep-brain stimulation for Parkinson’s disease,”New England Journal of Medicine, vol. 362, no. 22, pp. 2077–2091, 2010.

[3] J. A. Obeso, C. W. Olanow, M. C. Rodriguez-Oroz, P. Krack,R. Kumar, and A. E. Lang, “Deep-brain stimulation of thesubthalamic nucleus or the pars interna of the globus pallidusin Parkinson’s disease,” New England Journal of Medicine, vol.345, no. 13, pp. 956–963, 2001.

[4] C. W. Olanow, R. L. Watts, and W. C. Koller, “An algorithm(decision tree) for the management of Parkinson’s disease:treatment guidelines,” Neurology, vol. 56, no. 11, pp. S1–S88,2001.

[5] M. C. Rodriguez-Oroz, J. A. Obeso, A. E. Lang et al., “Bilat-eral deep brain stimulation in Parkinson’s disease: a multi-centre study with 4 years follow-up,” Brain, vol. 128, no. 10,pp. 2240–2249, 2005.



[8] J. A. Robichaud, K. D. Pfann, C. L. Comella, M. Brandabur,and D. M. Corcos, “Greater impairment of extension move-ments as compared to flexion movements in Parkinson’sdisease,” Experimental Brain Research, vol. 156, no. 2, pp.240–254, 2004.


[9] B. Durmus, O. Baysal, S. Altinayar, Z. Altay, Y. Ersoy, andC. Ozcan, “Lower extremity isokinetic muscle strength inpatients with Parkinson’s disease,” Journal of Clinical Neuro-science, vol. 17, no. 7, pp. 893–896, 2010.

[10] S. Kakinuma, H. Nogaki, B. Pramanik, and M. Morimatsu,“Muscle weakness in Parkinson’s disease: isokinetic study ofthe lower limbs,” European Neurology, vol. 39, no. 4, pp. 218–222, 1998.

[11] K. J. Bridgewater and M. H. Sharpe, “Trunk muscle perfor-mance in early Parkinson’s disease,” Physical Therapy, vol. 78,no. 6, pp. 566–576, 1998.

[12] P. Brown, D. M. Corcos, and J. C. Rothwell, “Does parkin-sonian action tremor contribute to muscle weakness inParkinson’s disease?” Brain, vol. 120, no. 3, pp. 401–408,1997.

[13] N. E. Allen, C. G. Canning, C. Sherrington, and V. S. C. Fung,“Bradykinesia, muscle Weakness and reduced muscle powerin Parkinson’s disease,” Movement Disorders, vol. 24, no. 9,pp. 1344–1351, 2009.

[14] W. Koller and S. Kase, “Muscle strength testing in Parkinson’sdisease,” European Neurology, vol. 25, no. 2, pp. 130–133,1986.

[15] B. K. Schilling, R. F. Pfeiffer, M. S. Ledoux, R. E. Karlage, R. J.Bloomer, and M. J. Falvo, “Effects of moderate-volume, high-load lower-body resistance training on strength and functionin persons with parkinson’s disease: a pilot study,” Parkinson’sDisease, Article ID 824734, 2010.

[16] D. M. Corcos, C. M. Chen, N. P. Quinn, J. McAuley, and J.C. Rothwell, “Strength in Parkinson’s disease: relationshipto rate of force generation and clinical status,” Annals ofNeurology, vol. 39, no. 1, pp. 79–88, 1996.

[17] A. E. Lang and A. M. Lozano, “Parkinson’s disease: second oftwo parts,” New England Journal of Medicine, vol. 339, no. 16,pp. 1130–1143, 1998.

[18] M. R. DeLong, “Primate models of movement disorders ofbasal ganglia origin,” Trends in Neurosciences, vol. 13, no. 7,pp. 281–285, 1990.

[19] T. Wichmann and M. R. DeLong, “Functional neuroanatomyof the basal ganglia in Parkinson’s disease,” Advances inNeurology, vol. 91, pp. 9–18, 2003.

[20] T. Wichmann and M. R. DeLong, “Anatomy and physiologyof the basal ganglia: relevance to Parkinson’s disease andrelated disorders,” Handbook of Clinical Neurology, vol. 83,pp. 1–18, 2007.

[21] M. DeLong and T. Wichmann, “Changing views of basalganglia circuits and circuit disorders,” Clinical EEG andNeuroscience, vol. 41, no. 2, pp. 61–67, 2010.

[22] H. Nogaki, S. Kakinuma, and M. Morimatsu, “Movementvelocity dependent muscle strength in Parkinson’s disease,”Acta Neurologica Scandinavica, vol. 99, no. 3, pp. 152–157,1999.

[23] K. D. Pfann, A. S. Buchman, C. L. Comella, and D. M. Corcos,“Control of movement distance in Parkinson’s disease,”Movement Disorders, vol. 16, no. 6, pp. 1048–1065, 2001.

[24] J. A. Robichaud, K. D. Pfann, C. L. Comella, and D.M. Corcos, “Effect of medication on EMG patterns inindividuals with Parkinson’s disease,” Movement Disorders,vol. 17, no. 5, pp. 950–960, 2002.

[25] J. J. J. Chen, T. Y. Sun, T. H. Lin, and T. S. Lin, “Spatio-temporal representation of multichannel EMG firing pat-terns and its clinical applications,” Medical Engineering andPhysics, vol. 19, no. 5, pp. 420–430, 1997.

[26] J. A. Robichaud, K. D. Pfann, D. E. Vaillancourt, C. L.Comella, and D. M. Corcos, “Force control and disease

severity in Parkinson’s disease,” Movement Disorders, vol. 20,no. 4, pp. 441–450, 2005.

[27] R. A. Miller, M. H. Thaut, G. C. McIntosh, and R. R. Rice,“Components of EMG symmetry and variability in parkin-sonian and healthy elderly gait,” Electroencephalographyand Clinical Neurophysiology—Electromyography and MotorControl, vol. 101, no. 1, pp. 1–7, 1996.

[28] J. Valls-Sole, A. Pascual-Leone, J. P. Brasil-Neto, A. Cam-marota, L. McShane, and M. Hallett, “Abnormal facilitationof the response to transcranial magnetic stimulation inpatients with Parkinson’s disease,” Neurology, vol. 44, no. 4,pp. 735–741, 1994.

[29] S. Salenius, S. Avikainen, S. Kaakkola, R. Hari, and P. Brown,“Defective cortical drive to muscle in Parkinson’s disease andits improvement with levodopa,” Brain, vol. 125, no. 3, pp.491–500, 2002.

[30] J. H. McAuley, D. M. Corcos, J. C. Rothwell, N. P. Quinn,and C. D. Marsden, “Levodopa reversible loss of the Piperfrequency oscillation component in Parkinson’s disease,”Journal of Neurology Neurosurgery and Psychiatry, vol. 70, no.4, pp. 471–476, 2001.

[31] L. Qi, J. M. Wakeling, A. Green, K. Lambrecht, and M.Ferguson-Pell, “Spectral properties of electromyographic andmechanomyographic signals during isometric ramp and stepcontractions in biceps brachii,” Journal of Electromyographyand Kinesiology, vol. 21, no. 1, pp. 128–135, 2011.

[32] D. Stoffers, J. L. W. Bosboom, J. B. Deijen, E. C. Wolters,H. W. Berendse, and C. J. Stam, “Slowing of oscillatorybrain activity is a stable characteristic of Parkinson’s diseasewithout dementia,” Brain, vol. 130, no. 7, pp. 1847–1860,2007.

[33] M. L. Petroni, G. Albani, V. Bicchiega et al., “Body composi-tion in advanced-stage Parkinson’s disease,” Acta Diabetolog-ica, vol. 40, no. 1, pp. S187–S190, 2003.

[34] H. S. Markus, A. M. Tomkins, and G. M. Stern, “Increasedprevalence of undernutrition in Parkinson’s disease and itsrelationship to clinical disease parameters,” Journal of NeuralTransmission—Parkinson’s Disease and Dementia Section, vol.5, no. 2, pp. 117–125, 1993.


[36] A. D. Speelman, B. P. Van De Warrenburg, M. Van Nimwe-gen, G. M. Petzinger, M. Munneke, and B. R. Bloem, “Howmight physical activity benefit patients with Parkinsondisease?” Nature Reviews Neurology, vol. 7, no. 9, pp. 528–534, 2011.

[37] D. E. Vaillancourt, J. Prodoehl, L. V. Metman, R. A. Bakay,and D. M. Corcos, “Effects of deep brain stimulation andmedication on bradykinesia and muscle activation in Parkin-son’s disease,” Brain, vol. 127, no. 3, pp. 491–504, 2004.

[38] D. E. Vaillancourt, J. Prodoehl, M. M. Sturman, R. A. E.Bakay, L. V. Metman, and D. M. Corcos, “Effects of deepbrain stimulation and medication on strength, bradykinesia,and electromyographic patterns of the ankle joint in Parkin-son’s disease,” Movement Disorders, vol. 21, no. 1, pp. 50–58,2006.

[39] M. M. Sturman, D. E. Vaillancourt, L. Verhagen Metman, R.A. E. Bakay, and D. M. Corcos, “Effects of five years of chronicSTN stimulation on muscle strength and movement speed,”Experimental Brain Research, vol. 205, pp. 435–443, 2010.

[40] O. R. Seynnes, M. de Boer, and M. V. Narici, “Early skeletalmuscle hypertrophy and architectural changes in response


to high-intensity resistance training,” Journal of AppliedPhysiology, vol. 102, no. 1, pp. 368–373, 2007.

[41] R. M. Enoka, “Neural adaptations with chronic physicalactivity,” Journal of Biomechanics, vol. 30, no. 5, pp. 447–455,1997.

[42] D. A. Gabriel, G. Kamen, and G. Frost, “Neural adaptationsto resistive exercise: mechanisms and recommendations fortraining practices,” Sports Medicine, vol. 36, no. 2, pp. 133–149, 2006.

[43] A. Holtermann, K. Roeleveld, B. Vereijken, and G. Ettema,“Changes in agonist EMG activation level during MVC can-not explain early strength improvement,” European Journal ofApplied Physiology, vol. 94, no. 5-6, pp. 593–601, 2005.

[44] J. Munn, R. D. Herbert, and S. C. Gandevia, “Contralateraleffects of unilateral resistance training: a meta-analysis,”Journal of Applied Physiology, vol. 96, no. 5, pp. 1861–1866,2004.

[45] T. Hortobagyi, N. J. Lambert, and J. P. Hill, “Greater crosseducation following training with muscle lengthening thanshortening,” Medicine and Science in Sports and Exercise, vol.29, no. 1, pp. 107–112, 1997.

[46] O. M. Rutherford and D. A. Jones, “The role of learningand coordination in strength training,” European Journal ofApplied Physiology and Occupational Physiology, vol. 55, no.1, pp. 100–105, 1986.

[47] T. J. Carroll, B. Barry, S. Riek, and R. G. Carson, “Resistancetraining enhances the stability of sensorimotor coordina-tion,” Proceedings of the Royal Society B: Biological Sciences,vol. 268, no. 1464, pp. 221–227, 2001.

[48] T. J. Carroll, S. Riek, and R. G. Carson, “The sites ofneural adaptation induced by resistance training in humans,”Journal of Physiology, vol. 544, no. 2, pp. 641–652, 2002.

[49] M. J. Falvo, E. J. Sirevaag, J. W. Rohrbaugh, and G. M.Earhart, “Resistance training induces supraspinal adapta-tions: evidence from movement-related cortical potentials,”European Journal of Applied Physiology, vol. 109, no. 5, pp.923–933, 2010.

[50] T. Liu-Ambrose, L. S. Nagamatsu, M. W. Voss, K. M. Khan,and T. C. Handy, “Resistance training and functional plas-ticity of theaging brain: a 12-month randomized controlledtrial,” Neurobiology of Aging. In press.

[51] K. Hakkinen and P. V. Komi, “Electromyographic changesduring strength training and detraining,” Medicine andScience in Sports and Exercise, vol. 15, no. 6, pp. 455–460,1983.

[52] A. R. Pucci, L. Griffin, and E. Cafarelli, “Maximal motorunit firing rates during isometric resistance training in men,”Experimental Physiology, vol. 91, no. 1, pp. 171–178, 2006.

[53] T. Moritani and H. A. DeVries, “Neural factors versus hyper-trophy in the time course of muscle strength gain,” AmericanJournal of Physical Medicine, vol. 58, no. 3, pp. 115–130, 1979.

[54] D. G. Sale, “Neural adaptation to resistance training,”Medicine and Science in Sports and Exercise, vol. 20, no. 5, pp.S135–S145, 1988.

[55] H. Akima, H. Takahashi, S. Y. Kuno et al., “Early phase adap-tations of muscle use and strength to isokinetic training,”Medicine and Science in Sports and Exercise, vol. 31, no. 4, pp.588–594, 1999.

[56] R. M. Enoka, “Muscle strength and its development. Newperspectives,” Sports Medicine, vol. 6, no. 3, pp. 146–168,1988.

[57] C. Patten, G. Kamen, and D. M. Rowland, “Adaptations inmaximal motor unit discharge rate to strength training in

young and older adults,” Muscle and Nerve, vol. 24, no. 4, pp.542–550, 2001.

[58] C. Del Balso and E. Cafarelli, “Adaptations in the activationof human skeletal muscle induced by short-term isometricresistance training,” Journal of Applied Physiology, vol. 103,no. 1, pp. 402–411, 2007.

[59] G. Kamen, “Aging, resistance training, and motor unit dis-charge behavior,” Canadian Journal of Applied Physiology, vol.30, no. 3, pp. 341–351, 2005.

[60] K. Hakkinen and P. V. Komi, “Changes in neuromuscularperformance in voluntary and reflex contraction duringstrength training in man,” International Journal of SportsMedicine, vol. 4, no. 4, pp. 282–288, 1983.

[61] B. Carolan and E. Cafarelli, “Adaptations in coactivation afterisometric resistance training,” Journal of Applied Physiology,vol. 73, no. 3, pp. 911–917, 1992.

[62] A. Holtermann, K. Roeleveld, M. Engstrøm, and T. Sand,“Enhanced H-reflex with resistance training is related toincreased rate of force development,” European Journal ofApplied Physiology, vol. 101, no. 3, pp. 301–312, 2007.

[63] J. T. Lemmer, G. F. Martel, D. E. Hurlbut, and B. F.Hurley, “Age and sex differentially affect regional changes inone repetition maximum strength,” Journal of Strength andConditioning Research, vol. 21, no. 3, pp. 731–737, 2007.

[64] G. F. Martel, S. M. Roth, F. M. Ivey et al., “Age and sex affecthuman muscle fibre adaptations to heavy-resistance strengthtraining,” Experimental Physiology, vol. 91, no. 2, pp. 457–464, 2006.


[66] G. M. Petzinger, J. P. Walsh, G. Akopian et al., “Effects oftreadmill exercise on dopaminergic transmission in the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-lesioned mousemodel of basal ganglia injury,” Journal of Neuroscience, vol.27, no. 20, pp. 5291–5300, 2007.

[67] M. B. Spraker, H. Yu, D. M. Corcos, and D. E. Vaillancourt,“Role of individual basal ganglia nuclei in force amplitudegeneration,” Journal of Neurophysiology, vol. 98, no. 2, pp.821–834, 2007.

[68] M. B. Spraker, J. Prodoehl, D. M. Corcos, C. L. Comella,and D. E. Vaillancourt, “Basal ganglia hypoactivity duringgrip force in drug naive Parkinson’s disease,” Human BrainMapping, vol. 31, no. 12, pp. 1928–1941, 2010.

[69] J. Prodoehl, M. Spraker, D. Corcos, C. Comella, and D.Vaillancourt, “Blood oxygenation level-dependent activationin basal ganglia nuclei relates to specific symptoms in de novoParkinson’s disease,” Movement Disorders, vol. 25, no. 13, pp.2035–2043, 2010.

[70] T. J. Carroll, V. S. Selvanayagam, S. Riek, and J. G. Semmler,“Neural adaptations to strength training: moving beyondtranscranial magnetic stimulation and reflex studies,” ActaPhysiologica, vol. 202, no. 2, pp. 119–140, 2011.


[72] N. E. Allen, C. Sherrington, C. G. Canning, and V. S. C.Fung, “Reduced muscle power is associated with slowerwalking velocity and falls in people with Parkinson’s disease,”


Parkinsonism and Related Disorders, vol. 16, no. 4, pp. 261–264, 2010.


[74] I. Reuter, M. Engelhardt, K. Stecker, and H. Baas, “Therapeu-tic value of exercise training in Parkinson’s disease,” Medicineand Science in Sports and Exercise, vol. 31, no. 11, pp. 1544–1549, 1999.

[75] K. Kuroda, K. Tatara, T. Takatorige, and F. Shinsho, “Effectof physical exercise on mortality in patients with Parkinson’sdisease,” Acta Neurologica Scandinavica, vol. 86, no. 1, pp. 55–59, 1992.


[77] C. J. Hass, M. A. Collins, and J. L. Juncos, “Resistance trainingwith creatine monohydrate improves upper-body strengthin patients with Parkinson disease: a randomized trial,”Neurorehabilitation and Neural Repair, vol. 21, no. 2, pp. 107–115, 2007.

[78] M. A. Hirsch, T. Toole, C. G. Maitland, and R. A. Rider, “Theeffects of balance training and high-intensity resistancetraining on persons with idiopathic Parkinson’s disease,”Archives of Physical Medicine and Rehabilitation, vol. 84, no.8, pp. 1109–1117, 2003.

[79] T. A. Scandalis, A. Bosak, J. C. Berliner, L. L. Helman, and M.R. Wells, “Resistance training and gait function in patientswith Parkinson’s disease,” American Journal of PhysicalMedicine and Rehabilitation, vol. 80, no. 1, pp. 38–43, 2001.

[80] S. L. Lindstedt, P. C. LaStayo, and T. E. Reich, “When activemuscles lengthen: properties and consequences of eccentriccontractions,” News in Physiological Sciences, vol. 16, no. 6,pp. 256–261, 2001.

[81] B. Bigland Ritchie and J. J. Woods, “Integrated electromyo-gram and oxygen uptake during positive and negative work,”Journal of Physiology, vol. 260, no. 2, pp. 267–277, 1976.

[82] A. Schrag, C. Sampaio, N. Counsell, and W. Poewe, “Minimalclinically important change on the Unified Parkinson’sDisease Rating Scale,” Movement Disorders, vol. 21, no. 8, pp.1200–1207, 2006.

[83] J. Jankovic and A. S. Kapadia, “Functional decline in Parkin-son disease,” Archives of Neurology, vol. 58, no. 10, pp. 1611–1615, 2001.

[84] Q. Xu, Y. Park, X. Huang et al., “Physical activities and futurerisk of Parkinson disease,” Neurology, vol. 75, no. 4, pp. 341–348, 2010.

[85] R. G. Brown and C. D. Marsden, “Cognitive function inParkinson’s disease: from description to theory,” Trends inNeurosciences, vol. 13, no. 1, pp. 21–29, 1990.

[86] J. A. Cooper, H. J. Sagar, N. Jordan, N. S. Harvey, andE. V. Sullivan, “Cognitive impairment in early, untreatedParkinson’s disease and its relationship to motor disability,”Brain, vol. 114, no. 5, pp. 2095–2122, 1991.

[87] A. J. Lees and E. Smith, “Cognitive deficits in the early stagesof Parkinson’s disease,” Brain, vol. 106, no. 2, pp. 257–270,1983.

[88] A. M. Owen, “Cognitive dysfunction in Parkinson’s disease:the role of frontostriatal circuitry,” Neuroscientist, vol. 10, no.6, pp. 525–537, 2004.

[89] A. E. Taylor, J. A. Saint-Cyr, and A. E. Lang, “Frontal lobedysfunction in Parkinson’s disease,” Brain, vol. 109, no. 5, pp.845–883, 1986.

[90] N. Sawamoto, M. Honda, T. Hanakawa et al., “Cognitiveslowing in Parkinson disease is accompanied by hypofunc-tioning of the striatum,” Neurology, vol. 68, no. 13, pp. 1062–1068, 2007.

[91] R. C. Cassilhas, V. A. R. Viana, V. Grassmann et al., “Theimpact of resistance exercise on the cognitive function of theelderly,” Medicine and Science in Sports and Exercise, vol. 39,no. 8, pp. 1401–1407, 2007.

[92] T. Liu-Ambrose, L. S. Nagamatsu, P. Graf, B. L. Beattie, M.C. Ashe, and T. C. Handy, “Resistance training and executivefunctions: a 12-month randomized controlled trial,” Archivesof Internal Medicine, vol. 170, no. 2, pp. 170–178, 2010.

[93] S. Colcombe and A. F. Kramer, “Fitness effects on thecognitive function of older adults: a meta-analytic study,”Psychological Science, vol. 14, no. 2, pp. 125–130, 2003.

[94] K. E. Cruise, R. S. Bucks, A. M. Loftus, R. U. Newton, R.Pegoraro, and M. G. Thomas, “Exercise and Parkinson’s:benefits for cognition and quality of life,” Acta NeurologicaScandinavica, vol. 123, no. 1, pp. 13–19, 2011.

[95] K. Tanaka, A. C. D. Quadros Jr., R. F. Santos, F. Stella, L.T. B. Gobbi, and S. Gobbi, “Benefits of physical exercise onexecutive functions in older people with Parkinson’s disease,”Brain and Cognition, vol. 69, no. 2, pp. 435–441, 2009.

[96] K. R. Vincent, R. W. Braith, T. Bottiglieri, H. K. Vincent,and D. T. Lowenthal, “Homocysteine, and lipoprotein levelsfollowing resistance training in older adults,” PreventiveCardiology, vol. 6, no. 4, pp. 197–203, 2003.

[97] S. E. Borst, D. V. De Hoyos, L. Garzarella et al., “Effects ofresistance training on insulin-like growth factor-I and IGFbinding proteins,” Medicine and Science in Sports and Exer-cise, vol. 33, no. 4, pp. 648–653, 2001.

[98] L. M. Garcia-Segura, M. A. Arevalo, and I. Azcoitia, “Interac-tions of estradiol and insulin-like growth factor-I signallingin the nervous system: new advances,” Progress in BrainResearch, vol. 181, pp. 251–272, 2010.

[99] S. Seshadri, A. Beiser, J. Selhub et al., “Plasma homocysteineas a risk factor for dementia and Alzheimer’s disease,” NewEngland Journal of Medicine, vol. 346, no. 7, pp. 476–483,2002.

[100] K. A. Jellinger and W. Paulus, “Clinicopathological correla-tions in Parkinsons-disease,” Clinical Neurology and Neuro-surgery, vol. 94, pp. S86–S88, 1992.

[101] W. J. Zetusky, J. Jankovic, and F. J. Pirozzolo, “The het-erogeneity of Parkinson’s disease: clinical and prognosticimplications,” Neurology, vol. 35, no. 4, pp. 522–526, 1985.

[102] A. H. Rajput, R. Pahwa, P. Pahwa, and A. Rajput, “Prognosticsignificance of the onset mode in parkinsonism,” Neurology,vol. 43, no. 4 I, pp. 829–830, 1993.

[103] G. Ransmayr, W. Poewe, and S. Plorer, “Prognostic impli-cations of the motor symptoms of Parkinson’s disease withrespect to clinical, computertomographic and psychometricparameters,” Journal of Neural Transmission—General Sec-tion, vol. 67, no. 1-2, pp. 1–14, 1986.


Review Article

Exercise and Motor Training in People with Parkinson’s Disease:A Systematic Review of Participant Characteristics,Intervention Delivery, Retention Rates, Adherence, andAdverse Events in Clinical Trials

Natalie E. Allen,1 Catherine Sherrington,2 Gayanthi D. Suriyarachchi,1

Serene S. Paul,1 Jooeun Song,1 and Colleen G. Canning1

1 Clinical and Rehabilitation Research Group, Faculty of Health Sciences, The University of Sydney, P.O. Box 170, Lidcombe,Sydney, NSW 1825, Australia

2 Musculoskeletal Division, The George Institute for Global Health, The University of Sydney, Sydney, NSW 2050, Australia

Correspondence should be addressed to Natalie E. Allen, [email protected]

Received 13 July 2011; Accepted 18 August 2011


Copyright © 2012 Natalie E. Allen et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

There is research evidence that exercise and motor training are beneficial for people with Parkinson’s disease (PD), and cliniciansseek to implement optimal programs. This paper summarizes important factors about the nature and reporting of randomizedcontrolled trials of exercise and/or motor training for people with PD which are likely to influence the translation of research intoclinical practice. Searches identified 53 relevant trials with 90 interventions conducted for an average duration of 8.3 (SD 4.2)weeks. Most interventions were fully supervised (74%) and conducted at a facility (79%). Retention rates were high with 69% ofinterventions retaining ≥85% of their participants; however adherence was infrequently reported, and 72% of trials did not reportadverse events. Overall, the labor-intensive nature of most interventions tested in these trials and the sparse reporting of adherenceand adverse events are likely to pose difficulties for therapists attempting to balance benefits and costs when selecting protocolsthat translate to sustainable clinical practice for people with PD.

1. Introduction

In recent years there have been an increasing number ofrandomized controlled trials assessing the effects of exerciseand/or motor training in people with Parkinson’s disease(PD). Overall, these trials support exercise and motortraining as beneficial in improving walking, balance, musclestrength, and the performance of functional tasks in peoplewith mild-to-moderate PD [1–11]. In order for findings fromthis research to be of general benefit to people with PD, ther-apists need to be able to translate the protocols used in theresearch into clinical practice [12].

Evidence-based practice aims to incorporate and applyhigh-quality clinical research findings in clinical policy andpractice [13, 14]. However, this can be a challenging task ashealth practitioners may find it difficult to assess, interpret,

and implement research evidence [13]. While evidence aboutbeneficial outcomes is paramount in therapists’ decisions toimplement a particular intervention, there are other factorsthat affect how the overall impact of the intervention is inter-preted and its potential for widespread clinical application[13–17]. For example, therapists need to consider how thecharacteristics of participants included in a trial may affecttheir decision regarding the applicability of the trial inter-vention with their patients [14]. It is likely that the way inwhich the intervention was applied in terms of its duration,level of supervision, delivery (i.e., individual versus group),and location (e.g., facilities and equipment required) willinfluence therapists’ decisions to implement that interven-tion. A research protocol that has been shown to be effectivemay not be implemented by therapists if they cannot provideadequate supervision over the required time frame or they do


not have access to necessary facilities or equipment. Finally,information regarding retention, adherence, and adverseevents is required so that therapists and patients can weigh upthe effectiveness of the intervention against its acceptabilityand any risks associated with implementation [14].

Therefore, in order to examine the information availableto guide the translation of research into clinical practice,we searched randomized controlled trials of exercise and/ormotor training for people with PD to determine the

(1) disease severity and cognitive status of the includedparticipants,

(2) duration, supervision, delivery, and location of theinterventions,

(3) rates of retention, adherence, and adverse events.

2. Methods

2.1. Data Sources and Searches. Randomized controlled trialsof exercise and/or motor training for people with PD wereidentified via database searches of MEDLINE, EMBASE,AMED, PsycINFO, the Cochrane Central Register of Con-trolled Trials, and CINAHL. The initial search was conductedin 2009, with a subsequent search conducted over 5 daysfrom the 7th of April, 2011. The electronic search strategyused has been previously reported [2]. The PhysiotherapyEvidence Database (PEDro; http://www.pedro.org.au/) wasalso searched, and the reference lists of previously publishedsystematic reviews [4, 5, 8, 9, 18–30] were checked for anytrials not identified with the database search.

2.2. Study Selection. Trials included were published ran-domized (or quasi randomized, i.e., not truly random butintended to produce similar groups, such as allocation by oddand even birth dates [31]) controlled trials of people withPD where at least one of the interventions was an ongoingprogram of exercise and/or motor training. All forms ofexercise (e.g., aerobic, strength, and treadmill walking) andmotor training (e.g., cueing and movement strategy training)were included. Whole-body vibration was not consideredto be exercise or motor training. Trials were excluded ifthe intervention was multidisciplinary or was primarilyoccupational therapy.

The eligibility of trials was determined in a two-stageprocess. Firstly, all trial titles and abstracts were screenedindependently by two investigators (N. E. Allen and G. D.Suriyarachchi). Trials were excluded if it was clear that theydid not meet the inclusion criteria. Secondly, the full articlewas obtained for the remaining trials and each trial wasassessed independently by at least two investigators (N. E.Allen, C. G. Canning or J. Song), using a standardized formcontaining the details of the inclusion criteria. Care was takento identify trials that had been reported in more than onejournal article. Where this occurred, the multiple articleswere counted as one trial and all articles were used to collectdata for that trial.

2.3. Data Extraction. A data collection form was developed,tested on five randomly selected trials and then modified

accordingly. All investigators were involved in data extrac-tion, and all data was double-checked by an investigator notinvolved in its initial extraction (N. E. Allen or J. Song).Discrepancies were resolved by discussion.

Information extracted from each trial included a descrip-tion of participants (including cognitive status), details ofthe exercise and motor training program and how it wasadministered, as well as details regarding retention rates,adherence to the intervention, and monitoring and reportingof adverse events. Retention was defined as the number ofparticipants who completed the trial (i.e., undertook thefirst or only post-intervention assessment excluding furtherfollow-up assessments) expressed as a percentage of thenumber of participants who began the trial. Adherence wasdefined as the number of intervention sessions attended byparticipants expressed as a percentage of the number ofintervention sessions prescribed [15].

3. Results

Searching identified 3,539 records, of which 53 trials involv-ing 1,940 participants were found to be eligible for inclusionin the paper (Figure 1) [32]. There were no disagreementsbetween reviewers regarding the inclusion of any articles. Thecharacteristics of the included trials [1, 3, 6, 7, 10, 11, 33–85]are summarised in Table 1.

3.1. Participant Characteristics. Forty (75%) of the reviewedtrials included participants with mild-to-moderate PD (i.e.,equivalent to Hoehn and Yahr stage I to III [86]). Seventrials (13%) included participants with mild-to-moderatelysevere PD (i.e., Hoehn and Yahr stage I to IV), while fourtrials (8%) included only participants with mild PD andtwo trials (4%) included only participants with moderatePD (Table 1). Most trials stipulated the cognitive status ofincluded participants. Twenty-nine trials (55%) used theMini-Mental State Examination [87] to screen potentialparticipants’ cognitive abilities, with the minimum score forinclusion varying between 20 and 28 out of the maximumof 30 [1, 3, 6, 7, 10, 11, 34, 36, 39, 41–43, 45, 47–50, 56,58, 60, 63–65, 69, 71, 76–78, 80, 85]. One trial (2%) [70]specified that participants required at least moderate scoreson the Neurobehavioural Cognitive Status Examination [88].Thirteen trials (25%) made a statement to the effect thatincluded participants had no dementia and/or reasonablecognition [35, 46, 51, 52, 57, 59, 62, 66, 67, 79, 81, 83, 84]. Tentrials (19%) did not give a clear indication of the participants’cognitive abilities [33, 37, 40, 53, 55, 68, 72, 74, 75, 82].

3.2. Exercise and/or Motor Training Program Characteristics.In the 53 trials, there were 90 intervention groups thatinvolved exercise and/or motor training (including twointervention groups for the cross-over trials where one inter-vention was a control [11, 42, 47]) (Table 1). Average inter-vention duration was 8.3 weeks (SD = 4.2, range = 2 to 26weeks), with 37 trials (70%) conducting an intervention of 10weeks or less. The total number of hours of intervention wasnot clearly reported in all studies (see Table 1); however, fromthe available data, an average of approximately 20 hours (SD


Records identified through

database searching

Scre

enin

gIn

clu

ded

Elig

ibili

tyId

enti

fica

tion

Additional records identified

through other sources

(n = 12)

Records after duplicates removed

Titles and abstracts screened Records excluded

Full-text articles assessed for

eligibility(n = 101)

Full-text articles excluded(n = 43)

19: Not randomized

9: Multidisciplinary or occupational

therapy intervention

7: Not exercise or motor training

2: Article not available

2: Protocol or commentary

2: Does not evaluate ongoing effects of

intervention

1: Participants did not have Parkinson’s

disease

1: Assessing a system of intervention

Trials included in review(n = 53 trials in 58 articles)

2 trials had 2 articles

article1 trial had 3 articles

2 trials were combined in a 3rd

(n = 3,527)

(n = 2,639) (n =2,538)

(n = 39)2,6

Figure 1: PRISMA flow diagram [32] showing flow of information through the review.

approximately 11, range = 4 to 65 hours) appears broadlyrepresentative of the included trials. Sixty-seven of the90 intervention groups (74%) involved full supervision ofexercise and/or motor training. Participants in 18 (27%) ofthe fully supervised intervention groups received one-on-one supervision and 20 (30%) received supervision insmall groups but the intervention delivery (one-on-one orsmall group supervision) was unclear in the remaining 29(43%) intervention groups. Participants in most interventiongroups (71; 79%) were required to attend a facility for all orthe majority of the intervention sessions.

3.3. Retention, Adherence, and Adverse Events. Retentionwas generally well reported and was high, with 62 (69%)of the 90 intervention groups retaining at least 85% ofparticipants (Table 1). Seventeen (32%) of the 53 includedtrials reported that at least one participant dropped out fora reason related to the intervention (Table 2). Difficulties

with transport and disinterest/poor adherence were the mostcommon intervention-related reasons for dropouts.

Overall, adherence and adverse events were infrequentlyreported in the included trials (Table 1). Adherence wasreported in some form in 26 (49%) of the included trials.However, 11 (42%) of these trials only reported adherencefor those participants who completed the intervention. Mosttrials (38; 72%) did not report monitoring for adverse events.Across the remaining 15 trials, 11 adverse events occurred(Table 1). Four participants from two separate trials [41, 80]experienced cardiac problems. Two of these participants,one from each group in a trial comparing physical therapywith and without mental practice [80], withdrew from thestudy. The two participants from the other trial [41] wereable to continue safely with treadmill training. Other adverseevents reported included a fall [81] and muscle cramps andtiredness [43] in trials involving cued overground walking,knee pain during a dancing program [52], muscle soreness


Ta

ble

1:C

har

acte

rist

ics

ofth

e53

incl

ude

dtr

ials

.

Firs

tau

thor

,yea

ran

din

terv

enti

onty

pe

Init

ialg

rou

psi

zes

Dis

ease

seve

rity

Loca

tion

and

deliv

ery

ofex

per

imen

tal

inte

rven

tion

Du

rati

onof

inte

rven

tion

(wee

ks)

Hou

rsof

inte

rven

tion

(app

rox)

Supe

rvis

ion

(%)

Ret

enti

on(%

)

Any

drop

outs

rela

ted

toin

terv

enti

on?

Adh

eren

ce(%

)

Adv

erse

even

tsoc

curr

edA

llen

2010

,M

ildto

Faci

lity,

grou

pex

erci

seO

R24

mod

erat

e+

hom

ein

divi

dual

2665

10%

87.5

%Y

70%

Nco

ntr

ol24

100%

Ash

burn

2007

,M

oder

ate

Hom

e,ex

erci

sean

d70

indi

vidu

al6

3318

%96

%Y

U(9

9%of

Sup

ervi

sed

exer

cise

)

Nst

rate

gyO

Rco

ntr

ol72

92%

Ber

gen

2002

,ae

robi

cex

erci

seO

R

Mild

Faci

lity,

NA

NR

NR

M4

deliv

ery

1632

100%

100%

un

clea

rco

ntr

ol4

100%

Bla

ckin

ton

2002

,U

Mild

toFa

cilit

y,U

NR

NR

Mex

erci

seO

R15

mod

erat

egr

oup

+h

ome

615

66%

53%

Yco

ntr

olto

tal

indi

vidu

alto

tal

Blo

omer

2008

,M

ildFa

cilit

y,re

sist

ance

8de

liver

y8

810

0%75

%N

UN

trai

nin

gO

Ru

ncl

ear

con

trol

888

%B

rau

n20

11,

Mild

toFa

cilit

y,U

NR

Mph

ysio

ther

apy

mod

erat

ely

indi

vidu

alor

6+

hom

e+

men

tal

25se

vere

grou

p+

hom

e,6

men

tal

U(1

00%

)‡88

%Y

(87%

)†‡

prac

tice

OR

indi

vidu

alpr

acti

ce

phys

ioth

erap

y+

rela

xati

on22

66

+h

ome

rela

xati

onU

(100

%)‡

82%

(88%

)†‡

Bri

dgew

ater

Mild

toFa

cilit

y,gr

oup

NA

NR

M19

96(a

lso

as[3

8]),

mod

erat

eae

robi

can

d13

1216

to22

100%

100%

95%

tru

nk

mu

scle

trai

nin

gO

Red

uca

tion

1312

410

0%U

Bu

rin

i200

6∗,

Mild

toFa

cilit

y,gr

oup

NR

Mae

robi

c13

mod

erat

e7

1710

0%85

%Y

87%

exer

cise

OR

Qig

ong

137

1710

0%85

%Y

88%

Cag

lar

2005

,M

ildto

Faci

lity,

grou

pN

AN

RM

exer

cise

OR

15m

oder

ate

+h

ome,

963

4%10

0%10

0%co

ntr

ol15

indi

vidu

al


Ta

ble

1:C

onti

nu

ed.

Firs

tau

thor

,yea

ran

din

terv

enti

onty

pe

Init

ialg

rou

psi

zes

Dis

ease

seve

rity

Loca

tion

and

deliv

ery

ofex

per

imen

tal

inte

rven

tion

Du

rati

onof

inte

rven

tion

(wee

ks)

Hou

rsof

inte

rven

tion

(app

rox)

Supe

rvis

ion

(%)

Ret

enti

on(%

)

Any

drop

outs

rela

ted

toin

terv

enti

on?

Adh

eren

ce(%

)

Adv

erse

even

tsoc

curr

edC

akit

2007

,M

ildto

Faci

lity,

NR

Part

ial—

trea

dmill

27m

oder

ate

deliv

ery

88

100%

78%

U2

min

orw

alki

ng

OR

un

clea

rca

rdia

cco

ntr

ol27

37%

Com

ella

UM

ildto

Faci

lity,

UN

RM

1994

∗ ,18

mod

erat

ede

liver

y89

%to

tal

phys

ioth

erap

yto

tal

un

clea

r4

1210

0%U

UO

Rco

ntr

olde

Bru

in20

10,

Mild

toH

ome,

wal

kw

ith

16m

oder

ate

indi

vidu

al13

19.5

0%69

%Y

100%

†Y—

1m

usi

cO

Rth

igh

con

trol

1765

%cr

ampi

ng

1ti

redn

ess

Der

eli2

010,

Mild

toN

RM

sup

ervi

sed

16m

oder

ate

Faci

lity,

1022

.510

0%94

%Y

100%

†

exer

cise

OR

hom

eex

erci

sein

divi

dual

;16

hom

e,in

divi

dual

1022

.54%

94%

NR

Dia

s20

05,

Mild

toFa

cilit

y,N

AN

RM

phys

ioth

erap

y8

mod

erat

ede

liver

y4

to10

2010

0%10

0%10

0%+

cues

OR

un

clea

rph

ysio

ther

apy

84

to10

2010

0%10

0%10

0%

Ebe

rsba

ch20

10,

Mild

tom

oder

ate

NR

NR

M

LSV

TB

IG20

Faci

lity,

indi

vidu

al;

416

+h

ome

XU

(100

%)‡

100%

NA

OR

Nor

dic

wal

kin

g20

faci

lity,

grou

p;8

16+

hom

eX

U(1

00%

)‡95

%U

OR

hom

eex

erci

se20

hom

e,in

divi

dual

4U

U95

%N

+al

ldid

hom

e,in

divi

dual

Elli

s20

05∗

Mild

toFa

cilit

y,gr

oup

NR

M(a

lso

as[4

4]),

mod

erat

eph

ysio

ther

apy

356

1810

0%91

%N

93%

OR

con

trol

3394

%N

Fish

er20

08,

Mild

Faci

lity,

NA

NR

Ntr

eadm

ill10

deliv

ery

818

100%

100%

wal

kin

gO

Ru

ncl

ear

phys

ioth

erap

y10

818

100%

100%

OR

con

trol

1010

0%


Ta

ble

1:C

onti

nu

ed.

Firs

tau

thor

,yea

ran

din

terv

enti

onty

pe

Init

ialg

rou

psi

zes

Dis

ease

seve

rity

Loca

tion

and

deliv

ery

ofex

per

imen

tal

inte

rven

tion

Du

rati

onof

inte

rven

tion

(wee

ks)

Hou

rsof

inte

rven

tion

(app

rox)

Supe

rvis

ion

(%)

Ret

enti

on(%

)

Any

drop

outs

rela

ted

toin

terv

enti

on?

Adh

eren

ce(%

)

Adv

erse

even

tsoc

curr

edFr

azzi

tta

2009

,M

oder

ate

Faci

lity,

NA

NR

NR

MTr

eadm

ill20

indi

vidu

al4

9.5

100%

100%

wal

kin

gw

ith

cues

OR

phys

ioth

erap

y20

49.

510

0%10

0%w

ith

cues

Gu

o20

09,

Mild

toFa

cilit

y,gr

oup

NR

NR

NR

Mph

ysio

ther

apy

23m

oder

ate

edu

cati

onan

din

divi

dual

812

ofth

erap

y10

0%91

%an

ded

uca

tion

ther

apy

OR

con

trol

2190

%H

ackn

ey20

07,

Mild

toFa

cilit

y,N

AN

RM

Tan

goO

R9

Mod

erat

egr

oup

1320

100%

100%

100%

exer

cise

1013

2010

0%10

0%10

0%H

ackn

ey20

08M

ildto

Faci

lity,

NR

M(a

lso

as[5

4]),

mod

erat

egr

oup

impl

ied

TaiC

hiO

R17

10to

1320

100%

76%

Y10

0%†

con

trol

1681

%H

ackn

ey20

09M

ildto

Faci

lity,

Part

ial—

(als

oas

[54]

),m

oder

ate

grou

p1

knee

Wal

tz/f

oxtr

ot19

10to

1320

100%

89%

Y10

0%†

pain

OR

tan

goO

R19

10to

1320

100%

74%

Y10

0%†

con

trol

2085

%H

ackn

ey20

10,

Mild

toFa

cilit

y,gr

oup

NR

Mpa

rtn

ered

19m

oder

ate

1020

100%

79%

Y10

0%†

tan

goO

Rn

onpa

rtn

ered

2010

2010

0%80

%Y

100%

†

tan

goH

ass

2007

,M

ildto

Faci

lity,

NA

NR

Y—1

resi

stan

ce10

mod

erat

ede

liver

y12

1210

0%10

0%m

usc

letr

ain

ing

+u

ncl

ear

sore

nes

ssu

pple

men

tan

d1

OR

Shou

lder

resi

stan

ce10

1212

100%

100%

pain

trai

nin

g+

plac

ebo


Ta

ble

1:C

onti

nu

ed.

Firs

tau

thor

,yea

ran

din

terv

enti

onty

pe

Init

ialg

rou

psi

zes

Dis

ease

seve

rity

Loca

tion

and

deliv

ery

ofex

per

imen

tal

inte

rven

tion

Du

rati

onof

inte

rven

tion

(wee

ks)

Hou

rsof

inte

rven

tion

(app

rox)

Supe

rvis

ion

(%)

Ret

enti

on(%

)

Any

drop

outs

rela

ted

toin

terv

enti

on?

Adh

eren

ce(%

)

Adv

erse

even

tsoc

curr

edH

irsc

h20

03,

Mild

toFa

cilit

y,ba

lan

cean

d9

mod

erat

ede

liver

y10

22.5

100%

67%

Y89

%†

Y—1

resi

stan

ceu

ncl

ear

ingu

inal

trai

nin

gO

Rh

ern

iaba

lan

ce9

1015

100%

100%

92%

trai

nin

gK

eus

2007

,M

ildto

Faci

lity,

NA

phys

ioth

erap

y14

mod

erat

ely

indi

vidu

al10

11to

1510

0%10

0%63

%N

OR

con

trol

13se

vere

92%

Ku

rtai

s20

08,

Mild

toFa

cilit

y,N

RN

RM

trea

dmill

13m

oder

ate

deliv

ery

612

+U

(100

%)‡

92%

Yw

alki

ng

OR

un

clea

r+

hom

eX

hom

e,co

ntr

ol14

indi

vidu

al86

%Le

hm

an20

05a ,

Mild

Faci

lity

impl

ied,

NA

NR

NR

Mw

alk

wit

h5

deliv

ery

un

clea

r2

510

0%10

0%ve

rbal

cues

OR

con

trol

610

0%

Mak

2008

,M

ildto

Faci

lity

NR

NR

Mcu

edsi

tto

21m

oder

ate

Impl

ied,

44

100%

90%

Ust

and

OR

deliv

ery

OR

exer

cise

21u

ncl

ear

46

100%

90%

Uco

ntr

ol18

78%

Mar

ches

eU

Mild

toFa

cilit

y,U

NR

NR

NR

M20

00,

20m

oder

ate

indi

vidu

al+

phys

ioth

erap

yto

tal

hom

e,6

18+

U(1

00%

) ‡O

Rin

divi

dual

hom

eX

phys

ioth

erap

y6

18+

U(1

00%

) ‡w

ith

cues

hom

eX

Miy

ai20

00∗ ,

Mild

toFa

cilit

y,N

AN

RN

RM

trea

dmill

5m

oder

ate

indi

vidu

al4

610

0%10

0%w

alki

ng

OR

phys

ioth

erap

y5

46

100%

100%

Miy

ai20

02,

Mild

toFa

cilit

y,N

RN

RM

trea

dmill

12m

oder

ate

indi

vidu

al4

610

0%92

%N

wal

kin

gO

Rph

ysio

ther

apy

124

610

0%75

%N


Ta

ble

1:C

onti

nu

ed.

Firs

tau

thor

,yea

ran

din

terv

enti

onty

pe

Init

ialg

rou

psi

zes

Dis

ease

seve

rity

Loca

tion

and

deliv

ery

ofex

per

imen

tal

inte

rven

tion

Du

rati

onof

inte

rven

tion

(wee

ks)

Hou

rsof

inte

rven

tion

(app

rox)

Supe

rvis

ion

(%)

Ret

enti

on(%

)

Any

drop

outs

rela

ted

toin

terv

enti

on?

Adh

eren

ce(%

)

Adv

erse

even

tsoc

curr

edM

orri

s20

09,

Mild

toFa

cilit

y,N

AN

RM

mov

emen

t14

mod

erat

ein

divi

dual

2U

pto

1210

0%10

0%88

%of

stra

tegi

esO

Rm

axim

um

sess

ion

sex

erci

ses

142

Up

to12

100%

100%

81%

ofm

axim

um

sess

ion

sM

ulle

r19

97,

Mild

toLo

cati

onan

dN

RN

RN

RM

beh

avio

ura

l15

mod

erat

ede

liver

y10

3010

0%U

ther

apy

OR

un

clea

rn

onsp

ecifi

c14

1030

100%

Uex

erci

ses

and

info

rmat

ion

Nie

uwbo

er20

07∗

(als

oas

[61,

73])

,M

ildto

mod

erat

eH

ome,

indi

vidu

alPa

rtia

l—n

ofa

llsw

eari

ng

acti

vity

mon

itor

cuei

ng

763

4.5

100%

99%

N10

0%tr

ain

ing

OR

con

trol

7710

0%Pa

cch

etti

Mild

toFa

cilit

y,gr

oup

UN

RN

RN

RM

2000

,m

oder

ate

phys

ioth

erap

y16

1319

.510

0%O

Rm

usi

cth

erap

y16

1326

100%

Palm

er19

86,

Mild

toFa

cilit

y,gr

oup

NA

NR

NR

Mex

erci

seO

R7

mod

erat

ely

1236

100%

100%

seat

edka

rate

7se

vere

1236

100%

100%

Pelo

sin

2010

,M

ildto

Faci

lity,

NA

NR

NR

Mph

ysio

ther

apy

9m

oder

ate

deliv

ery

412

100%

100%

wit

hFO

Gu

ncl

ear

stra

tegi

esO

Rph

ysio

ther

apy

94

1210

0%10

0%P

rota

s20

05,

Mild

toFa

cilit

y,N

AN

RM

trea

dmill

9m

oder

ate

indi

vidu

al8

2410

0%10

0%10

0%w

alki

ng

and

step

trai

nin

gO

Rco

ntr

ol9

100%


Ta

ble

1:C

onti

nu

ed.

Firs

tau

thor

,yea

ran

din

terv

enti

onty

pe

Init

ialg

rou

psi

zes

Dis

ease

seve

rity

Loca

tion

and

deliv

ery

ofex

per

imen

tal

inte

rven

tion

Du

rati

onof

inte

rven

tion

(wee

ks)

Hou

rsof

inte

rven

tion

(app

rox)

Supe

rvis

ion

(%)

Ret

enti

on(%

)

Any

drop

outs

rela

ted

toin

terv

enti

on?

Adh

eren

ce(%

)

Adv

erse

even

tsoc

curr

edQ

utu

budd

inM

ildto

Faci

lity,

NR

M20

07,

mod

erat

ede

liver

yco

mpu

teri

zed

12u

ncl

ear

+8

1820

%75

%U

Udy

nam

ich

ome,

post

uro

grap

hyin

divi

dual

OR

phys

ioth

erap

y10

818

20%

60%

UU

Rid

gel2

009,

Mild

toFa

cilit

y,N

AN

RN

RM

forc

edcy

clin

g5

mod

erat

ede

liver

y8

2410

0%10

0%O

Ru

ncl

ear

self

-pac

ed5

824

100%

100%

cycl

ing

Sage

2009

,M

ildto

Faci

lity,

NR

Mae

robi

c17

mod

erat

egr

oup

10to

1218

100%

76%

Y86

.8%†

exer

cise

OR

SAFE

xO

R21

10to

1220

to34

100%

86%

Y92

.9%†

con

trol

1510

0%Sa

ge20

10,

Mild

toFa

cilit

y,N

AN

RM

SAFE

xO

R15

mod

erat

egr

oup

1236

100%

100%

100%

non

-SA

FEx

1412

3610

0%10

0%10

0%Sc

hen

kman

Mild

toLo

cati

on19

98,

mod

erat

eu

ncl

ear,

exer

cise

OR

27in

divi

dual

10to

1322

.5to

3010

0%85

%U

100%†

Nco

ntr

ol24

96%

Sch

mit

z-M

ildto

Faci

lity

grou

pN

RN

RM

Hu

bsch

2006

,m

oder

atel

y+

hom

eQ

igon

gO

R32

seve

rein

divi

dual

2416

+h

ome

X66

%91

%Y

con

trol

2487

.5%

Sman

ia20

10,

Mod

erat

eFa

cilit

y,N

RN

RM

bala

nce

33to

indi

vidu

al7

17.5

100%

85%

Ytr

ain

ing

OR

mod

erat

ely

gen

eral

31se

vere

717

.510

0%87

%Y

exer

cise

sSt

allib

rass

Mild

toFa

cilit

y,N

RM

2002

,m

oder

ate

indi

vidu

alA

lexa

nde

r32

1216

100%

91%

Y99

%†

tech

niq

ue

OR

mas

sage

OR

3112

1693

%97

%†

con

trol

3010

0%


Ta

ble

1:C

onti

nu

ed.

Firs

tau

thor

,yea

ran

din

terv

enti

onty

pe

Init

ialg

rou

psi

zes

Dis

ease

seve

rity

Loca

tion

and

deliv

ery

ofex

per

imen

tal

inte

rven

tion

Du

rati

onof

inte

rven

tion

(wee

ks)

Hou

rsof

inte

rven

tion

(app

rox)

Supe

rvis

ion

(%)

Ret

enti

on(%

)

Any

drop

outs

rela

ted

toin

terv

enti

on?

Adh

eren

ce(%

)

Adv

erse

even

tsoc

curr

edTa

mir

2007

,M

ildto

Faci

lity,

NR

Part

ial—

exer

cise

+im

ager

y12

mod

erat

ede

liver

y12

24+

U(1

00%

)‡92

%U

2ca

rdia

cO

Ru

ncl

ear

+h

ome

Xpr

oble

ms

exer

cise

11h

ome,

indi

vidu

al12

24+

hom

eX

U(1

00%

)‡91

%U

Th

aut

1996

,M

ildto

Hom

e,N

Aw

alk

wit

h15

mod

erat

ein

divi

dual

310

.514

%10

0%10

0%Y—

1fa

llau

dito

rycu

esO

Rw

alk

wit

hou

t11

310

.514

%10

0%10

0%cu

esO

Rco

ntr

ol11

100%

Tool

e20

00,

Mild

toFa

cilit

y,N

RN

RM

exer

cise

OR

6m

oder

atel

yde

liver

y10

3010

0%67

%N

con

trol

5se

vere

un

clea

r60

%To

ole

2005

,U

Mild

toFa

cilit

y,N

RN

R99

%N

RM

trea

dmill

23m

oder

atel

yde

liver

y6

610

0%ov

eral

lw

alki

ng

wit

hto

tal

seve

reu

ncl

ear

body

wei

ght

supp

ort

OR

trea

dmill

66

100%

wal

kin

gw

ith

wei

ghts

OR

trea

dmill

66

100%

wal

kin

gYa

ng

2010

,M

ildto

Faci

lity,

dow

nh

ill16

mod

erat

ede

liver

y4

610

0%94

%Y

100%

†N

trea

dmill

un

clea

rw

alki

ng

OR

conv

enti

onal

174

610

0%88

%Y

100%

†N

ther

apy

You

sefi

2009

,M

ildto

Faci

lity,

NA

NR

NR

Mex

erci

seO

R12

mod

erat

egr

oup

1040

100%

100%

edu

cati

on12

1040

100%

Supe

rvis

ion

:th

en

um

ber

ofin

terv

enti

onse

ssio

ns

supe

rvis

edex

pres

sed

asa

perc

enta

geof

the

nu

mbe

rof

sess

ion

spr

escr

ibed

;ret

enti

on:t

he

nu

mbe

rof

part

icip

ants

wh

oco

mpl

eted

the

tria

l(i.e

.,u

nde

rtoo

ka

post

-in

terv

enti

onas

sess

men

tbu

tex

clu

din

gfo

llow

-up)

expr

esse

das

ape

rcen

tage

ofth

en

um

ber

ofpa

rtic

ipan

tsw

ho

bega

nth

etr

ial;

adh

eren

ce:t

he

nu

mbe

rof

inte

rven

tion

sess

ion

spa

rtic

ipan

tsat

ten

ded

expr

esse

das

ap

erce

nta

geof

the

nu

mbe

rof

inte

rven

tion

sess

ion

spr

escr

ibed

;Y:y

es;N

:no;

NA

:not

appl

icab

le;U

:un

clea

r—in

suffi

cien

tin

form

atio

nto

cate

gori

ze;N

R:n

otre

por

ted;

NR

M:n

otre

por

ted

tobe

mon

itor

ed;X

:ex

erci

se;F

OG

:fre

ezin

gof

gait

;SA

FEx:

sen

sory

atte

nti

onfo

cuse

dex

erci

se;∗

cros

s-ov

ertr

ial;†

data

only

for

part

icip

ants

wh

oco

mpl

eted

the

tria

l;‡da

taon

lyfo

rth

efa

cilit

yco

mpo

nen

tof

the

inte

rven

tion

;a part

2of

tria

lon

ly.


Table 2: Dropout reasons when related to the intervention.

First author and year Dropout reasonNumber ofparticipants

Allen 2010 Did not want to do the intervention 1

Ashburn 2007 Falls (but not during intervention) 1

Blackinton 2002 Safety concerns 1

Braun 2011 Imagery too confronting 1

Burini 2006Poor adherence to exercise; 2

back pain 1

de Bruin 2010 No access to necessary equipment 1

Dereli 2010 Did not want to do the intervention 1

Hackney 2008 (also as [54])Exercise not intense enough; 1

transport problems 2

Hackney 2009 (also as [54])Knee pain; 1


Hackney 2010Travel distance; 2

classes too fatiguing; 1

lack of interest 1

Hirsch 2003 Inguinal hernia 1

Kurtais 2008 Poor adherence to exercise 1

Sage 2009 Time commitment 4

Schmitz-Hubsch 2006Uncomfortable in the group; 1

uncomfortable with Qigong 1

Smania 2010 Uncooperative 4

Stallibrass 2002 Could not travel 1

Yang 2010Low motivation; 1


and shoulder pain [56] following resistance training, anda hernia [57] subsequent to muscle strength assessment.

4. Discussion

A substantial number of randomized controlled trials ofexercise and/or motor training for people with PD havebeen published. However, the nature and reporting of thesetrials are likely to provide challenges for therapists aiming toimplement the interventions into clinical practice [17]. Mosttrials involved only cognitively intact participants with mild-to-moderate PD. Trials tended to be of short duration, highlysupervised, and conducted at a facility. Furthermore, thereports for many trials were lacking important details, withadherence and adverse events particularly being inadequatelyreported.

On the whole, trials included in this paper included onlyparticipants with mild-to-moderate PD who were withoutsignificant cognitive impairment. Including only these typesof participants not only makes it easier to conduct trialsof exercise and motor training interventions but also aidsinterpretation of the results. However, cognitive impairment

is now recognised as a common problem in PD, with over80% of people with PD ultimately developing dementia [89].Further work is needed to determine the effectiveness ofexercise and motor training in people with more severecognitive impairment and/or more advanced disease.

Most of the reviewed trials were of short duration, highlysupervised, and facility based (Table 1). Interventions lastedan average of around two months. Seventy-four percentof the intervention groups were fully supervised, with noreported expectation for participants to undertake unsu-pervised exercise. Furthermore, 79% of intervention groupswere mainly conducted at a facility such as a hospital oruniversity. Such brief, highly supervised interventions con-ducted in controlled environments are likely to improve theadherence of participants to exercise and motor trainingprograms and to ensure that interventions are being per-formed optimally. In this regard, these trials are useful andimportant for determining the short-term efficacy of anintervention. However, given that PD is a long-term, neu-rodegenerative condition, the capacity of therapists andpatients to sustain the intervention over the long termneeds to be considered. Furthermore, such brief and highlysupervised interventions are costly and less likely to give


information about the effectiveness of the intervention whenimplemented into usual practice [13, 17]. For example, therequirement for participants to travel to a facility was acommon reason for withdrawal from the included trials(Table 2). Moreover, the neurodegenerative nature of PDand the limited resources available to healthcare systemsmean that such labor-intensive programs are unlikely tobe sustained or afforded by most health-care providers.Additionally, as PD is a progressive disease it is importantthat people with PD are empowered to self-manage theirdisease to some extent [90, 91]. To this end, trials ofmore pragmatic and sustainable exercise and motor traininginterventions, with the potential for direct translation intoclinical practice and including cost-effectiveness analysis, areneeded.

The likely adherence to an exercise and motor trainingprogram is an important factor to consider when prescribingsuch a program for an individual with PD. Adherence to theintervention was reported in less than half of the includedtrials, and some reports of adherence are artificially elevatedby including only those participants who completed thetrial (Table 1). Some trials were able to effectively maximiseadherence by providing a flexible timeframe for participantsto complete the intervention [46, 51, 52, 74, 76] and soallow participants more options in fitting their exerciseand/or motor training program around their daily lives. Thispragmatic approach is likely to more closely reflect therapyattendance patterns and is therefore likely to be helpful fortherapists considering translating the research into theirclinical practice.

Given the importance of adherence to exercise and motortraining programs, strategies to promote adherence in peoplewith PD need to be considered. Providing a high level ofsupervision seems likely to promote adherence in the shortterm, as it may enhance participants’ commitment to theprogram. However, a Cochrane review comparing home andcentre-based exercise programs for older adults found that,in the long term, participants were more likely to adhere tohome-based programs [92]. Furthermore, the reviewersnoted a trend toward more sustained improvements in thehome-based than in the centre-based programs and sug-gested that this was attributable to the higher adherence inhome-based programs. In the present paper, three of theincluded trials report high levels of adherence with mini-mally supervised home-based programs [40, 43, 81]. Com-mon to all three of these trials was a requirement forparticipants to keep a daily record of what exercise/motortraining they had performed. It seems likely that this simplestrategy assisted in promoting adherence in these trials.Other strategies with the potential to improve adherence insustainable, minimally supervised trials, such as participantinvolvement in goal setting [93, 94], flexibility to allowprograms to be modified for individuals [1, 91, 93, 94], andintermittent followup [91, 94], warrant exploration.

The issue of adverse events was inadequately addressed inthe trials included in this paper, with only 15 trials reportingmonitoring for adverse events. In these 15 trials, 11 adverseevents were reported, most of which were minor in nature(Table 1). However, when discussing and planning exercise

and motor training options with people with PD, therapistsneed to be informed not only about the effectiveness of agiven intervention but also about the nature and likelihoodof any potential adverse events [95]. Similarly poor reportingof adverse events was found in a recent Cochrane review ofprogressive resistance training for older adults [95]. Notably,the Cochrane review found that adverse events were morelikely to be detected in trials that used a clear definition ofadverse events than in trials which did not use a definition.In the same way, the use of a definition for adverse eventsis likely to improve the assessment and reporting of adverseevents in trials of exercise and motor training for people withPD.

This paper has examined several factors in the nature andreporting of trials of exercise and/or motor training whichare likely to influence the way research is applied by therapistsin clinical practice. However, this paper did not addresswhether or not trial protocols were reported in sufficientdetail to allow therapists to emulate the research interventionin the clinic. This detailed reporting of trial interventions iscritical in enabling research to be clinically applied [96]. Theability of many journals to provide online material whichsupplements the published article will aid the provision ofsuch details despite the necessary word count limitationsplaced on authors.

5. Conclusions

Clinicians seeking to use research to inform their clinicalpractice rely heavily on the design and reporting of ran-domized controlled trials to reach their decisions. However,the nature and reporting of trials of exercise and/or motortraining for people with PD are likely to provide challengesfor therapists aiming to implement the interventions intoclinical practice. The short duration, highly supervised andfacility-based nature of many of the interventions, coupledwith the tendency to include only cognitively-intact partic-ipants with mild-to-moderate disease, mean that findingsmay not generalise when therapists set out to apply them inthe long-term management of people with PD. Infrequentreporting of adherence and adverse events compounds thisproblem and makes cost-benefit balancing more difficult. Itis recommended that these issues be taken into account inthe design and reporting of future trials.

References


[2] N. E. Allen, C. Sherrington, S. S. Paul, and C. G. Canning,“Balance and falls in Parkinson’s disease: a meta-analysis ofthe effect of exercise and motor training,” Movement Disorders,vol. 26, no. 9, pp. 1605–1615, 2001.



[4] C. J. de Goede, S. H. J. Keus, G. Kwakkel, and R. C. Wagenaar,“The effects of physical therapy in Parkinson’s disease:a research synthesis,” Archives of Physical Medicine andRehabilitation, vol. 82, no. 4, pp. 509–515, 2001.


[6] G. Ebersbach, A. Ebersbach, D. Edler et al., “Comparingexercise in Parkinson’s disease—the Berlin BIG study,”Movement Disorders, vol. 25, no. 12, pp. 1902–1908, 2010.

[7] G. Ebersbach, A. Ebersbach, D. Edler et al., “Comparingexercie in Parkinson’s disease—the Berlin LSVT�BIG study,”Movement Disorders, vol. 25, no. 14, p. 2478, 2010.

[8] V. A. Goodwin, S. H. Richards, R. S. Taylor, A. H. Taylor, andJ. L. Campbell, “The effectiveness of exercise interventionsfor people with Parkinson’s disease: a systematic reviewand meta-analysis,” Movement Disorders, vol. 23, no. 5, pp.631–640, 2008.

[9] J. Mehrholz, R. Friis, J. Kugler, S. Twork, A. Storch, andM. Pohl, “Treadmill training for patients with Parkinson’sdisease,” Cochrane Database of Systematic Reviews, no. 1,Article ID CD007830, 2010.

[10] M. E. Morris, R. Iansek, and B. Kirkwood, “A randomizedcontrolled trial of movement strategies compared withexercise for people with Parkinson’s disease,” MovementDisorders, vol. 24, no. 1, pp. 64–71, 2009.

[11] A. Nieuwboer, G. Kwakkel, L. Rochester et al., “Cueingtraining in the home improves gait-related mobility inParkinson’s disease: The RESCUE trial,” Journal of Neurology,Neurosurgery and Psychiatry, vol. 78, no. 2, pp. 134–140, 2007.

[12] R. C. Brownson and E. Jones, “Bridging the gap: translatingresearch into policy and practice,” Preventive Medicine, vol.49, no. 4, pp. 313–315, 2009.

[13] B. Haynes and A. Haines, “Getting research findings into prac-tice. Barriers and bridges to evidence based clinical practice,”British Medical Journal, vol. 317, no. 7153, pp. 273–276, 1998.

[14] R. Herbert, G. Jamtvedt, J. Mead, and K. Birger Hagen,Practical Evidence-Based Physiotherapy, Elsevier, Edinburgh,UK, 2005.

[15] E. V. Cyarto, W. J. Brown, and A. L. Marshall, “Retention,adherence and compliance: important considerations forhome- and group-based resistance training programs forolder adults,” Journal of Science and Medicine in Sport, vol. 9,no. 5, pp. 402–412, 2006.

[16] L. Quinn, M. Busse, H. Khalil, S. Richardson, A. Rosser,and H. Morris, “Client and therapist views on exerciseprogrammes for early-mid stage Parkinson’s disease andHuntington’s disease,” Disability and Rehabilitation, vol. 32,no. 11, pp. 917–928, 2010.

[17] T. A. Sheldon, G. H. Guyatt, and A. Haines, “Getting researchfindings into practice. When to act on the evidence,” BritishMedical Journal, vol. 317, no. 7151, pp. 139–142, 1998.

[18] A. M. Crizzle and I. J. Newhouse, “Is physical exercisebeneficial for persons with Parkinson’s disease?” ClinicalJournal of Sport Medicine, vol. 16, no. 5, pp. 422–425, 2006.

[19] K. H. Deane, D. Jones, C. Ellis-Hill, C. E. Clarke, E. D.Playford, and Y. Ben-Shlomo, “A comparison of physiotherapytechniques for patients with Parkinson’s disease,” CochraneDatabase of Systematic Reviews, no. 1, Article ID CD002815,2001.

[20] K. Deane, D. Jones, E. Playford, Y. Ben-Shlomo, and C.Clarke, “Physiotherapy versus placebo or no intervention in

Parkinson’s disease,” Cochrane Database of Systematic Reviews,no. 3, Article ID CD002817, 2001.

[21] K. H. O. Deane, C. Ellis-Hill, D. Jones et al., “Systematicreview of paramedical therapies for Parkinson’s disease,”Movement Disorders, vol. 17, no. 5, pp. 984–991, 2002.

[22] H. Gage and L. Storey, “Rehabilitation for Parkinson’sdisease: a systematic review of available evidence,” ClinicalRehabilitation, vol. 18, no. 5, pp. 463–482, 2004.

[23] M. Jobges, H. Spittler-Schneiders, C. I. E. Renner, and H.Hummelsheim, “Clinical relevance of rehabilitation programsfor Parkinson’s disease. I: non-symptom-specific therapeuticapproaches,” Parkinsonism and Related Disorders, vol. 13, no.4, pp. 195–202, 2007.

[24] E. M. Jobges, H. Spittler-Schneiders, C. I. E. Renner, andH. Hummelsheim, “Clinical relevance of rehabilitationprograms for patients with idiopathic Parkinson syndrome.II: symptom-specific therapeutic approaches,” Parkinsonismand Related Disorders, vol. 13, no. 4, pp. 203–213, 2007.


[26] G. Kwakkel, C. J. T. de Goede, and E. E. H. van Wegen,“Impact of physical therapy for Parkinson’s disease: a criticalreview of the literature,” Parkinsonism and Related Disorders,vol. 13, supplement 3, pp. S478–S487, 2007.

[27] I. Lim, E. van Wegen, C. de Goede et al., “Effects of externalrhythmical cueing on gait in patients with Parkinson’s disease:a systematic review,” Clinical Rehabilitation, vol. 19, no. 7, pp.695–713, 2005.

[28] A. Nieuwboer, “Cueing for freezing of gait in patients withParkinson’s disease: a rehabilitation perspective,” MovementDisorders, vol. 23, supplement 2, pp. S475–S481, 2008.

[29] T. C. Rubinstein, N. Giladi, and J. M. Hausdorff, “The powerof cueing to circumvent dopamine deficits: a review of physicaltherapy treatment of gait disturbances in Parkinson’s disease,”Movement Disorders, vol. 17, no. 6, pp. 1148–1160, 2002.

[30] O. Suchowersky, G. Gronseth, J. Perlmutter, S. Reich,T. Zesiewicz, and W. J. Weiner, “Practice parameter:Neuroprotective strategies and alternative therapies forParkinson disease (an evidence-based review),” Neurology,vol. 66, no. 7, pp. 976–982, 2006.

[31] J. P. T. Higgins and S. Green, Cochrane Handbook forSystematic Reviews of Interventions Version 5.1.0, TheCochrane Collaboration, 2009.

[32] A. Liberati, D. G. Altman, J. Tetzlaff et al., “The PRISMAstatement for reporting systematic reviews and meta-analysesof studies that evaluate health care interventions: explanationand elaboration,” Journal of Clinical Epidemiology, vol. 62, no.10, pp. e1–e34, 2009.

[33] J. L. Bergen, T. Toole, R. G. Elliott Jr., B. Wallace, K. Robinson,and C. G. Maitland, “Aerobic exercise intervention improvesaerobic capacity and movement initiation in Parkinson’sdisease patients,” NeuroRehabilitation, vol. 17, no. 2, pp.161–168, 2002.

[34] M. Blackinton, L. Summerall, and K. Waguespack, “Tertiaryprevention in Parkinson disease: results from a preliminarystudy,” Neurology Report, vol. 26, no. 3, pp. 160–165, 2002.

[35] R. J. Bloomer, B. K. Schilling, R. E. Karlage, M. S. Ledoux, R. F.Pfeiffer, and J. Callegari, “Effect of resistance training on bloodoxidative stress in Parkinson disease,” Medicine and Science inSports and Exercise, vol. 40, no. 8, pp. 1385–1389, 2008.


[36] S. Braun, A. Beurskens, M. Kleynen, J. Schols, and D. Wade,“Rehabilitation with mental practice has similar effects onmobility as rehabilitation with relaxation in people withParkinson’s disease: a multicentre randomised trial,” Journalof Physiotherapy, vol. 57, no. 1, pp. 27–34, 2011.

[37] K. J. Bridgewater and M. H. Sharpe, “Aerobic exercise andearly parkinson’s disease,” Neurorehabilitation and NeuralRepair, vol. 10, no. 4, pp. 233–241, 1996.

[38] K. J. Bridgewater and M. H. Sharpe, “Trunk muscle trainingand early Parkinson’s disease,” Physiotherapy Theory andPractice, vol. 13, no. 2, pp. 139–153, 1997.

[39] D. Burini, B. Farabollini, S. Iacucci et al., “A randomisedcontrolled cross-over trial of aerobic training versus Qigongin advanced Parkinson’s disease,” Europa Medicophysica, vol.42, no. 3, pp. 231–238, 2006.

[40] A. T. Caglar, H. N. Gurses, F. K. Mutluay, and G. Kiziltan,“Effects of home exercises on motor performance in patientswith Parkinson’s disease,” Clinical Rehabilitation, vol. 19, no.8, pp. 870–877, 2005.

[41] B. D. Cakit, M. Saracoglu, H. Genc, H. R. Erdem, and L. Inan,“The effects of incremental speed-dependent treadmill train-ing on postural instability and fear of falling in Parkinson’s dis-ease,” Clinical Rehabilitation, vol. 21, no. 8, pp. 698–705, 2007.

[42] C. L. Comella, G. T. Stebbins, N. Brown-Toms, and C. G.Goetz, “Physical therapy and Parkinson’s disease: a controlledclinical trial,” Neurology, vol. 44, no. 3, part 1, pp. 376–378,1994.

[43] N. de Bruin, J. B. Doan, G. Turnbull et al., “Walking withmusic is a safe and viable tool for gait training in parkinson’sdisease: the effect of a 13-week feasibility study on single anddual task walking,” Parkinson’s Disease, vol. 2010, Article ID483530, 9 pages, 2010.

[44] C. J. T. de Goede, T. Ellis, and R. C. Wagenaar, “Effecten vaneen fysiotherapie-groepsbehandeling Parkinson-patienten:een cross-over trial,” Nederlands Tijdschrift voor Fysiotherapie,vol. 114, no. 3, pp. 78–82, 2004.

[45] E. E. Dereli and A. Yaliman, “Comparison of the effectsof a physiotherapist-supervised exercise programme anda self-supervised exercise programme on quality of life inpatients with Parkinson’s disease,” Clinical Rehabilitation, vol.24, no. 4, pp. 352–362, 2010.

[46] N. Dias, D. Fraga, E. Cacho, and T. Oberg, “Training of gaitwith visual cues in the patient with Parkinson’s disease,”Fisioterapia em Movimento, vol. 18, no. 4, pp. 43–51, 2005.

[47] T. Ellis, C. J. de Goede, R. G. Feldman, E. C. Wolters,G. Kwakkel, and R. C. Wagenaar, “Efficacy of a physicaltherapy program in patients with Parkinson’s disease: arandomized controlled trial,” Archives of Physical Medicineand Rehabilitation, vol. 86, no. 4, pp. 626–632, 2005.


[49] G. Frazzitta, R. Maestri, D. Uccellini, G. Bertotti, and P. Abelli,“Rehabilitation treatment of gait in patients with Parkinson’sdisease with freezing: a comparison between two physicaltherapy protocols using visual and auditory cues with orwithout treadmill training,” Movement Disorders, vol. 24, no.8, pp. 1139–1143, 2009.

[50] L. Guo, Y. Jiang, H. Yatsuya, Y. Yoshida, and J. Sakamoto,“Group education with personal rehabilitation for idiopathicParkinson’s disease,” Canadian Journal of NeurologicalSciences, vol. 36, no. 1, pp. 51–59, 2009.

[51] M. E. Hackney and G. M. Earhart, “Tai Chi improves balanceand mobility in people with Parkinson disease,” Gait andPosture, vol. 28, no. 3, pp. 456–460, 2008.

[52] M. E. Hackney and G. M. Earhart, “Effects of dance onmovement control in Parkinson’s disease: a comparisonof Argentine tango and American ballroom,” Journal ofRehabilitation Medicine, vol. 41, no. 6, pp. 475–481, 2009.

[53] M. E. Hackney and G. M. Earhart, “Effects of dance on gaitand balance in Parkinsons disease: a comparison of partneredand nonpartnered dance movement,” Neurorehabilitation andNeural Repair, vol. 24, no. 4, pp. 384–392, 2010.

[54] M. E. Hackney and G. M. Earhart, “Health-related quality oflife and alternative forms of exercise in Parkinson disease,”Parkinsonism and Related Disorders, vol. 15, no. 9, pp.644–648, 2009.

[55] M. E. Hackney, S. Kantorovich, R. Levin, and G. M. Earhart,“Effects of tango on functional mobility in Parkinson’s disease:a preliminary study,” Journal of Neurologic Physical Therapy,vol. 31, no. 4, pp. 173–179, 2007.

[56] C. J. Hass, M. A. Collins, and J. L. Juncos, “Resistance trainingwith creatine monohydrate improves upper-body strengthin patients with Parkinson disease: a randomized trial,”Neurorehabilitation and Neural Repair, vol. 21, no. 2, pp.107–115, 2007.

[57] M. A. Hirsch, T. Toole, C. G. Maitland, and R. A. Rider,“The effects of balance training and high-intensity resistancetraining on persons with idiopathic Parkinson’s disease,”Archives of Physical Medicine and Rehabilitation, vol. 84, no. 8,pp. 1109–1117, 2003.

[58] S. H. J. Keus, B. R. Bloem, J. J. van Hilten, A. Ashburn, and M.Munneke, “Effectiveness of physiotherapy in Parkinson’s dis-ease: the feasibility of a randomised controlled trial,” Parkin-sonism and Related Disorders, vol. 13, no. 2, pp. 115–121, 2007.

[59] Y. Kurtais, S. Kutlay, B. S. Tur, H. Gok, and C. Akbostanci,“Does treadmill training improve lower-extremity tasks inParkinson disease? A randomized controlled trial,” ClinicalJournal of Sport Medicine, vol. 18, no. 3, pp. 289–291, 2008.

[60] D. A. Lehman, T. Toole, D. Lofald, and M. A. Hirsch, “Trainingwith verbal instructional cues results in near-term improve-ment of gait in people with Parkinson disease,” Journal ofNeurologic Physical Therapy, vol. 29, no. 1, pp. 2–8, 2005.

[61] I. Lim, E. van Wegen, D. Jones et al., “Does cueing trainingimprove physical activity in patients with parkinson’sdisease?” Neurorehabilitation and Neural Repair, vol. 24, no. 5,pp. 469–477, 2010.



[64] I. Miyai, Y. Fujimoto, Y. Ueda et al., “Treadmill trainingwith body weight support: its effect on Parkinson’s disease,”Archives of Physical Medicine and Rehabilitation, vol. 81, no. 7,pp. 849–852, 2000.

[65] I. Miyai, Y. Fujimoto, H. Yamamoto et al., “Long-term effectof body weight-supported treadmill training in Parkinson’sdisease: a randomized controlled trial,” Archives of PhysicalMedicine and Rehabilitation, vol. 83, no. 10, pp. 1370–1373,2002.


[66] V. Muller, B. Mohr, R. Rosin, F. Pulvermuller, F. Muller, andN. Birbaumer, “Short-term effects of behavioral treatmenton movement initiation and postural control in Parkinson’sdisease: a controlled clinical study,” Movement Disorders, vol.12, no. 3, pp. 306–314, 1997.

[67] C. Pacchetti, F. Mancini, R. Aglieri, C. Fundaro, E. Martignoni,and G. Nappi, “Active music therapy in Parkinson’s disease: anintegrative method for motor and emotional rehabilitation,”Psychosomatic Medicine, vol. 62, no. 3, pp. 386–393, 2000.

[68] S. S. Palmer, J. A. Mortimer, D. D. Webster, R. Bistevins, andG. L. Dickinson, “Exercise therapy for Parkinson’s disease,”Archives of Physical Medicine and Rehabilitation, vol. 67, no.10, pp. 741–745, 1986.

[69] E. Pelosin, L. Avanzino, M. Bove, P. Stramesi, A. Nieuwboer,and G. Abbruzzese, “Action observation improves freezing ofgait in patients with Parkinson’s disease,” Neurorehabilitationand Neural Repair, vol. 24, no. 8, pp. 746–752, 2010.

[70] E. J. Protas, K. Mitchell, A. Williams, H. Qureshy, K. Caroline,and E. C. Lai, “Gait and step training to reduce falls inParkinson’s disease,” NeuroRehabilitation, vol. 20, no. 3, pp.183–190, 2005.

[71] A. A. Qutubuddin, D. X. Cifu, P. Armistead-Jehle, W. Carne, T.E. McGuirk, and M. S. Baron, “A comparison of computerizeddynamic posturography therapy to standard balance physicaltherapy in individuals with Parkinson’s disease: a pilot study,”NeuroRehabilitation, vol. 22, no. 4, pp. 261–265, 2007.

[72] A. L. Ridgel, J. L. Vitek, and J. L. Alberts, “Forced, notvoluntary, exercise improves motor function in Parkinson’sdisease patients,” Neurorehabilitation and Neural Repair, vol.23, no. 6, pp. 600–608, 2009.

[73] L. Rochester, K. Baker, V. Hetherington et al., “Evidencefor motor learning in Parkinson’s disease: acquisition,automaticity and retention of cued gait performance aftertraining with external rhythmical cues,” Brain Research, vol.1319, pp. 103–111, 2010.

[74] M. D. Sage and Q. J. Almeida, “Symptom and gait changesafter sensory attention focused exercise vs aerobic training inParkinson’s disease,” Movement Disorders, vol. 24, no. 8, pp.1132–1138, 2009.

[75] M. D. Sage and Q. J. Almeida, “A positive influence of visionon motor symptoms during sensory attention focused exercisefor Parkinson’s disease,” Movement Disorders, vol. 25, no. 1,pp. 64–69, 2010.


[77] T. Schmitz-Hubsch, D. Pyfer, K. Kielwein, R. Fimmers, T.Klockgether, and U. Wullner, “Qigong exercise for the symp-toms of Parkinson’s disease: a randomized, controlled pilotstudy,” Movement Disorders, vol. 21, no. 4, pp. 543–548, 2006.

[78] N. Smania, E. Corato, M. Tinazzi et al., “Effect of balancetraining on postural instability in patients with idiopathicparkinsong’s disease,” Neurorehabilitation and Neural Repair,vol. 24, no. 9, pp. 826–834, 2010.

[79] C. Stallibrass, P. Sissons, and C. Chalmers, “Randomizedcontrolled trial of the Alexander Technique for idiopathicParkinson’s disease,” Clinical Rehabilitation, vol. 16, no. 7, pp.695–708, 2002.

[80] R. Tamir, R. Dickstein, and M. Huberman, “Integration ofmotor imagery and physical practice in group treatmentapplied to subjects with Parkinson’s disease,” Neurorehabilita-tion and Neural Repair, vol. 21, no. 1, pp. 68–75, 2007.

[81] M. H. Thaut, G. C. McIntosh, R. R. Rice, R. A. Miller, J.Rathbun, and J. M. Brault, “Rhythmic auditory stimulationin gait training for Parkinson’s disease patients,” MovementDisorders, vol. 11, no. 2, pp. 193–200, 1996.

[82] T. Toole, M. A. Hirsch, A. Forkink, D. A. Lehman, and C.G. Maitland, “The effects of a balance and strength trainingprogram on equilibrium in Parkinsonism: a preliminarystudy,” NeuroRehabilitation, vol. 14, no. 3, pp. 165–174, 2000.

[83] T. Toole, C. G. Maitland, E. Warren, M. F. Hubmann, and L.Panton, “The effects of loading and unloading treadmill walk-ing on balance, gait, fall risk, and daily function in Parkinson-ism,” NeuroRehabilitation, vol. 20, no. 4, pp. 307–322, 2005.

[84] Y. R. Yang, Y. Y. Lee, S. J. Cheng, and R. Y. Wang, “Downhillwalking training in individuals with parkinson’s disease: arandomized controlled trial,” American Journal of PhysicalMedicine and Rehabilitation, vol. 89, no. 9, pp. 706–714, 2010.

[85] B. Yousefi, V. Tadibi, A. F. Khoei, and A. Montazeri, “Exercisetherapy, quality of life, and activities of daily living in patientswith Parkinson disease: a small scale quasi-randomised trial,”Trials, vol. 10, article 67, 2009.

[86] M. M. Hoehn and M. D. Yahr, “Parkinsonism: onset,progression and mortality,” Neurology, vol. 17, no. 5, pp.427–442, 1967.

[87] M. F. Folstein, S. E. Folstein, and P. R. McHugh, “’Mini mentalstate’. A practical method for grading the cognitive state ofpatients for the clinician,” Journal of Psychiatric Research, vol.12, no. 3, pp. 189–198, 1975.

[88] M. E. Oehlert, S. D. Hass, M. R. Freeman, M. D. Williams, J.J. Ryan, and S. W. Sumerall, “The neurobehavioral cognitivestatus examination: accuracy of the ’Screen-Metric’ approachin a clinical sample,” Journal of Clinical Psychology, vol. 53, no.7, pp. 733–737, 1997.

[89] M. A. Hely, W. G. J. Reid, M. A. Adena, G. M. Halliday, andJ. G. L. Morris, “The Sydney Multicenter Study of Parkinson’sdisease: the inevitability of dementia at 20 years,” MovementDisorders, vol. 23, no. 6, pp. 837–844, 2008.

[90] G. Guyatt, D. Cook, and B. Haynes, “Evidence based medicinehas come a long way,” British Medical Journal, vol. 329, no.7473, pp. 990–991, 2004.

[91] M. E. Morris, C. L. Martin, and M. L. Schenkman, “Stridingout with Parkinson disease: evidence-based physical therapyfor gait disorders,” Physical Therapy, vol. 90, no. 2, pp.280–288, 2010.

[92] N. L. Ashworth, K. E. Chad, E. L. Harrison, B. A. Reeder, andS. C. Marshall, “Home versus center based physical activityprograms in older adults,” Cochrane Database of SystematicReviews, no. 1, Article ID CD004017, 2005.

[93] M. Schenkman, D. Hall, R. Kumar, and W. M. Kohrt,“Endurance exercise training to improve economy ofmovement of people with Parkinson disease: three casereports,” Physical Therapy, vol. 88, no. 1, pp. 63–76, 2008.

[94] M. van Nimwegen, A. D. Speelman, K. Smulders et al., “Designand baseline characteristics of the ParkFit study, a randomizedcontrolled trial evaluating the effectiveness of a multifacetedbehavioral program to increase physical activity in Parkinsonpatients,” BMC Neurology, vol. 10, article 70, 2010.

[95] C. J. Liu and N. K. Latham, “Progressive resistance strengthtraining for improving physical function in older adults,”Cochrane Database of Systematic Reviews, no. 3, Article IDCD002759, 2009.

[96] D. T. Wade, R. J. E. M. Smeets, and J. A. Verbunt, “Research inrehabilitation medicine: methodological challenges,” Journalof Clinical Epidemiology, vol. 63, no. 7, pp. 699–704, 2010.


Research Article

Effectiveness of an Inpatient Movement Disorders Program forPatients with Atypical Parkinsonism

Anna D. Hohler,1, 2 Jyeming M. Tsao,1 Douglas I. Katz,1, 2 T. Joy DiPiero,1, 2

Christina L. Hehl,2 Alissa Leonard,2 Valerie Allen,2 Maura Gardner,2 Heidi Phenix,2

Marie Saint-Hilaire,1, 2 and Terry Ellis2, 3

1 Department of Neurology, Boston University School of Medicine, 720 Albany Street, Suite 7B, Boston, MA 02118, USA2 Braintree Rehabilitation Hospital, Braintree, MA 250 Pond Street, Braintree, MA 02184, USA3 Department of Physical Therapy, College of Health & Rehabilitation Sciences, Boston University, 635 Commonwealth Avenue,Boston, MA 02215, USA

Correspondence should be addressed to Anna D. Hohler, [email protected]



Copyright © 2012 Anna D. Hohler et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

This paper investigated the effectiveness of an inpatient movement disorders program for patients with atypical parkinsonism, whotypically respond poorly to pharmacologic intervention and are challenging to rehabilitate as outpatients. Ninety-one patientswith atypical parkinsonism participated in an inpatient movement disorders program. Patients received physical, occupational,and speech therapy for 3 hours/day, 5 to 7 days/week, and pharmacologic adjustments based on daily observation and data.Differences between admission and discharge scores were analyzed for the functional independence measure (FIM), timed upand go test (TUG), two-minute walk test (TMW), Berg balance scale (BBS) and finger tapping test (FT), and all showed significantimprovement on discharge (P > .001). Clinically significant improvements in total FIM score were evident in 74% of the patients.Results were similar for ten patients whose medications were not adjusted. Patients with atypical parkinsonism benefit from aninpatient interdisciplinary movement disorders program to improve functional status.

1. Introduction

Atypical parkinsonism is used to describe disorders charac-terized by parkinsonism—tremor, rigidity, akinesia, and pos-tural instability—but not caused by Parkinson’s disease (PD).These disorders often include other prominent features. Theterm includes progressive supranuclear palsy (PSP), multiplesystem atrophy (MSA), Lewy body dementia (LBD), corti-cobasal degeneration (CBD), vascular parkinsonism, drug-induced parkinsonism, and parkinsonism secondary to in-fection and other causes. PSP is characterized by parkinson-ism along with a supranuclear vertical gaze palsy and earlyonset of balance problems and falls. The hallmark features ofMSA include parkinsonism, autonomic instability, and cere-bellar and corticospinal deficits. LBD has similar pathologyto PD; however, accumulation of Lewy bodies in areas out-side the substantia nigra leads to hallucinations, cognitive

impairment and dementia prior to the onset of parkinson-ism. Features of CBD include asymmetric parkinsonism,apraxia, alien limb phenomenon, aphasia, and sensory de-ficits. Vascular parkinsonism is due to lacunar infarcts in thebasal ganglia and can be distinguished from PD by an abruptonset or stepwise deterioration and development of parkin-sonism and evidence of neurovascular disease. Drug-inducedparkinsonism is often due to antipsychotics or antiemeticsand usually resolves with cessation of the offending drug[1, 2].

Treatment of PD involves medications that increase dop-amine in the basal ganglia. However, there has been less suc-cess with pharmacologic treatment in atypical parkinsonism[1, 3–5]. Previous studies have shown physical therapy to bean effective adjunctive treatment in patients with PD [6, 7]but there have only been a handful of case reports and studiesinvestigating the efficacy of nonpharmacologic therapy inatypical parkinsonism patients.


Case reports and studies have shown subjective and ob-jective improvements in gait, balance, and patient safety inpatients with PSP [8–12] and in a patient with mixed CBDand PSP [13]. Similar improvements in gait, balance, trans-fers, and stability were seen in case reports of physical therapyintervention in patients with MSA [14, 15]. Timed up andgo, functional reach test, 360-degree turn, and 50 foot timedwalk are examples of some of the improved objective meas-ures. Another case report showed improvement in activitiesof daily life (ADLs) and finger manipulation after repetitivefinger exercises in a patient with CBD [16]. A small pilot ran-domized controlled trial showed significant improvement inMSA patients after receiving individualized outpatient occu-pational therapy [17].

Regarding intensive inpatient programs, a prior study byEllis et al. investigated the effectiveness of an inpatient re-habilitation program for people with PD. In the study, medi-cation was adjusted and interdisciplinary rehabilitation pro-gram was provided to optimize patients’ functional ability.Significant improvements were noted in all outcome mea-sures. Patients who did not have changes made to their medi-cations also showed significant improvements in total,motor, and cognitive functional independence measure(FIM) scores [18].

The purpose of this study was to investigate the effec-tiveness of an inpatient movement disorders program in im-proving functional status for patients with atypical parkin-sonism and to determine whether or not these findingswere clinically meaningful. We hypothesized that people withatypical parkinsonism would show statistically and clinicallysignificant improvements in functional status after partici-pating in such a program.

2. Methods

2.1. Design and Subjects. A pretest-posttest design was usedto determine the effectiveness of a movement disorder pro-gram for patients admitted to an inpatient rehabilitation hos-pital with the diagnosis of atypical parkinsonism. Patientswere admitted from home, acute care facilities, skilled nurs-ing facilities, or assisted living between January 2004 andAugust 2008. They carried diagnoses, determined by a neuro-logist specializing in movement disorders, which fall underthe term “atypical parkinsonism” as described above. Theywere at least 18 years old and were given a Hoehn and Yahrstage I to V for classification of PD. A total of 91 subjects wereadmitted to the program. Baseline characteristics are listed inTable 1.

2.2. Outcome Measures. Primary outcomes were FIM totalscore. Secondary outcomes included FIM motor score, FIMcognitive score, 2-minute walk test (TMW), Timed “up andgo” test (TUG) Berg balance score, and finger tapping test(FT).

The FIM is a widely used 18-item assessment of disabilityamong inpatient rehabilitation patients. The FIM measuresability to perform basic life activities, such as self-care, sphin-cter control, transfers, locomotion, communication, and

Table 1: Patient baseline characteristics.

Characteristic

Age—y, mean (SD) n = 91 76.6 (7.7)

Disease duration—y, mean (SD) n = 91 5.7 (4.4)

Sex no. (%) men/women n = 91 53 (58.2)/38 (41.8)

Race no. (%) white n = 91 84 (92.3)

Education no. (%) ≤ bachelor’s/>bachelor’s n = 75

51 (56)/24 (26.4)

Hoehn and Yahr stage no. (%) n = 85

I 0 (0)

II 0 (0)

III 26 (28.6)

IV 43 (47.3)

V 16 (17.6)

Length of stay—d, mean (SD) 24.4 (11.9)

Diagnosis no. (%) n = 91a

Vascular parkinsonism 25 (27.5)

MSA 19 (20.9)

PSP 4 (4.4)

Medication related 2 (2.2)

LBD 1 (1.1)

CBD 1 (1.1)

Toxin 1 (1.1)

Unknown 38 (41.8)aMSA: multiple system atrophy, PSP: progressive supranuclear palsy, LBD:

Lewy body dementia, and CBD: corticobasal dementia.

social cognition. Each item is scored on a scale from 1 to 7,in which 1 is patient requires total assistance to complete thetask and 7 is complete independence. The FIM can be dividedinto 2 sections: motor (13 items) and cognitive (5 items). Ithas been shown to have good reliability and validity [19].

The TMW is performed by asking subjects to walk as faras they can in 2 minutes. Patients with PD have been shownto cover less distance than age-matched controls [20].

The TUG assesses a patient’s ability to transfer from sitt-ing to standing, ambulate, and make a turn. Patients are tim-ed while rising from a chair, walking 3 m, turning, walkingback to the chair, and sitting down. It has been shown to havehigh interrater reliability for subjects with PD [21]. Subjectswere allowed to use an assistive device if necessary for theTMW and the TUG.

The BBS is a 14-item scale assessing balance while sitting,standing, turning, and reaching forward. Items are ratedfrom 0 to 4, with 0 meaning the subject needs assistance or isunable to perform the task and 4 meaning the subject canperform the task safely and independently. It has been shownto be reliable and valid in patients with PD [22, 23]. Minimaldetectable change (MCD) has been found to be +/−6 pointsamong patients who have suffered a stroke [24] and +/−5points in patients with PD [25].

The FT is a timed test useful in assessing the impact ofbradykinesia on rapid alternating movements of the upperextremity. Two buttons are attached to a counter 30 cm apart.


Subjects are asked to alternate tapping each button with theirleft hand for one minute. The sum of the taps is the score forthat hand. The test is repeated with the right hand. The FThas been shown to have good validity and reliability and isable to distinguish normal subjects from those with PD [26].

2.3. Intervention. A multidisciplinary team consisting ofneurologists specializing in movement disorders and neuro-rehabilitation, physical therapists, occupational therapists,speech-language pathologists, nurses, and case managersprovided a comprehensive rehabilitation program for pa-tients admitted to the hospital.

All outcome measures were obtained at admission anddischarge, as well as daily measurements of TMW and TUGat the peak and troughs of medication cycles. The same ther-apists administered interventions and outcome measures.Weekly rounds were conducted to allow the whole team toevaluate the data and discuss the patients’ status. Decisionsregarding changes to medications or rehabilitation interven-tions were made at this time. Subjects’ responses to medica-tion adjustments were discussed further at weekly movementdisorder meetings. Adjustments were made to subjects’ ther-apy and medications during the entire length of stay at thehospital. Medication adjustments included increases or de-creases in Parkinson’s disease medications to optimize peakperformance. Those medications included carbidopa/ lev-odopa, monoamine oxidase inhibitors, catechol-o methyl-transferase inhibitors and amantadine.

Subjects received individually tailored physical, occupa-tional, and speech therapy for a minimum of 3 hours per dayfor 5 to 7 days per week. Therapy was provided on an indi-vidual and group basis. Interventions included external cue-ing to improve gait speed, step length and cadence [27–30],cognitive movement strategies during task-based training toimprove mobility, balance and transfers [6, 31–33], resistiveexercises [34], exercises for joint mobility [35, 36], andspeech therapy to improve voice volume and clarity [37, 38].A more detailed description of the intervention is providedin Table 2.

2.4. Data Analysis. Means, standard deviations, and fre-quency distributions were calculated for subjects’ baselinecharacteristics, length of stay, and disposition. The efficacyof the intervention was evaluated by comparing admissionand discharge mean scores for each of the outcome measures.Two-tailed paired t tests were conducted with an alpha levelset at 0.05. A Bonferroni-adjusted type I error rate (α = 0.007)was applied to all t tests. Results were calculated for patientswho received rehabilitation along with PD medication ad-justments and for patients who received rehabilitation only.Clinically significant improvement was determined based ona total FIM score change of ≥22, which has been associatedwith the minimal clinically important difference in peoplewho have suffered a stroke [39]. In this study a change fromadmission of more than 22 was considered to be clinicallymeaningful.

3. Results

3.1. Subjects. Ninety-one subjects with atypical parkinson-ism underwent rehabilitation therapy. Average age at admis-sion was 76.5 years (SD 0.81), and they had been carryingthe diagnosis of parkinsonism for an average of 5.7 years(SD 0.55). Of the 91 subjects with atypical parkinsonism, 25(27.5%) had vascular parkinsonism, 19 (20.9%) had MSA, 4(4.4%) had PSP, and the remaining cases were either medi-cation related, due to LBD, CBD, toxin exposure or were un-known. Eighty-five of the subjects were Hoehn and Yahrstages III-V. Six subjects were not evaluated using Hoehnand Yahr (Table 1). The rehabilitation team made changesto 81 subjects’ medications while receiving physical therapy.Ten subjects underwent physical rehabilitation only with nochanges to their medications. Length of stay varied from oneto six weeks with an average stay of 2.5 weeks.

3.2. Outcomes for All Patients. Statistically significant im-provements were made in all outcome measures over thecourse of the rehabilitation program (Table 3). Total FIMscore increased 29.5 points (95% CI = 26.4–32.5). Addition-ally, motor FIM improved 25.9 points (95% CI = 23.4–28.5)and cognitive FIM improved 3.5 points (95% CI = 2.6–4.4).TUG decreased by 39.4 seconds (95% CI = 20.6–58.2), TMWlengthened by 63.5 feet (95% CI = 44.3–82.9), and Bergbalance scale scores improved 7.5 points (95% CI = 4.3–10.6). Left and right finger tapping improved by 11.5 taps(95% CI = 6.7–16.1) and 10.8 taps (95% CI = 5.8–16.1), res-pectively.

Previous studies have shown the minimal clinically im-portant difference of total FIM to be 22 [39]. Using this cut-off, sixty-five (74%) patients made clinically meaningful im-provements in total FIM.

3.3. Outcomes for Patients Receiving Rehabilitation Only andNo Changes to Medication. Statistically significant improve-ments were made in all but left finger tapping for the 10patients who received rehabilitation only (Table 4). Analysisshowed an improvement of 32.1 (95% CI = 22.8–41.3) fortotal FIM, 28.6 (95% CI = 19.8–37.3) for motor FIM, and3.5 (95% CI = 1.7–5.2) for cognitive FIM. TUG decreased by52 seconds on average (95% CI = 13.1–91.0), TMW lengthincreased by 76 feet (95% CI = 27.4–124.5), and right fingertapping improved by 19.6 taps (95% CI = 3.4–25.7). Leftfinger tapping increased on average 7.8 taps, but results werenot statistically significant (95% CI = −9.7–25.4).

4. Discussion

This study investigated the effectiveness of an interdisci-plinary inpatient rehabilitation program for patients withatypical parkinsonism. Our results showed improvements intotal FIM, motor FIM, cognitive FIM, TMW, TUG, BBS, andleft and right FT. Among patients who received rehabilitationonly, without changes to their medication regimens, statisti-cally significant improvement in all but left FT was observed.Clinically meaningful improvement, defined by a change in


Table 2: Description of interventions.

Functional training

Rolling from supine position to sitting position and from sitting position to supine position.

Transferring from sitting position to standing position, from chair to bed, and from chair to toilet.

Dressing and grooming.

Balance: reactive and anticipatory within functional contexts.

Gait training

Walking with external auditory cues from a metronome to optimize gait speed and cadence;increasing cadence by 10% over baseline and progressing until cadence approaches normal oruntil subject reaches maximum capacity.

Reducing freezing (context specific: doorways, thresholds, and narrow spaces) with visual cues inthe form of lines on the floor from tape or laser beams.

Reducing freezing (context specific: doorways, thresholds, and narrow spaces) with visual cues inthe form of lines on the floor from tape or laser beams.

Improving adaptation (various walking surfaces, obstacles in the environment, starting andstopping, and turning head while walking).

Curb negotiation and stair climbing.

Range of motion, flexibility,strengthening exercises

Range of motion (increase trunk extension and rotation).

Stretching (hip flexor, hamstring, and gastrocnemius muscles).

Strengthening (trunk and hip postural muscles and knee and ankle extensor muscles).

Speech exercisesExercises to improve vocal rate control.

Exercises to improve phonation.

Table 3: All patients.

Measure (n)a Admission mean (SD) Discharge mean (SD) Change (95% CI)

Total FIM (88) 41.3 (15.7) 70.8 (21.4) 29.5 (28.4, 32.5)

Motor FIM (88) 23.6 (11) 49.6 (16.6) 25.9 (23.4, 28.5)

Cognitive FIM (87) 17.7 (6.1) 21.2 (5.8) 3.5 (2.6, 4.4)

TUG (60) 81.5 s (89.4) 42.0 s (46.9) −39.4 s (−20.6, −28.2)

TMW (60) 138.9 ft (76.9) 202.5 ft (96.9) 63.5 ft (44.2, 82.9)

BBS (28) 22 (12.2) 29.5 (13.4) 7.5 (4.3, 10.6)

Left FT (51) 60.2 (23.3) 71.7 (24.8) 11.5 (6.7, 16.1)

Right FT (50) 68.3 (27) 79.3 (30) 10.9 (5.8, 16.1)aFIM: functional independence measure, TUG: timed up and go, TMW: 2-minute walk, BBS: berg balance scale, and FT: finger tap.

total FIM of greater than 22, was also observed in 74% ofpatients. These results imply that patients with atypical par-kinsonism can show significant improvement in functionafter receiving intensive inpatient multidisciplinary therapyincluding rehabilitative and pharmacologic interventions.

Patients with atypical parkinsonism are difficult to man-age on an outpatient basis. The complexity of their symp-toms, the added cognitive and autonomic deficits, the poorresponse to most PD pharmacological agents, and the rela-tively rapid decline in status contribute to the challenges inmanaging these patients particularly as the disease progress-es. This study highlights the benefits of an interdisciplinaryrehabilitation program in addition to medication adjust-ments in an inpatient setting, where patients could be ob-served over a 24-hour period, 7 days per week by health careprofessionals with expertise in movement disorders. Objec-tive measures taken daily during peak and troughs of medi-cation cycle allowed an objective, systematic assessmentof function. This data was used to guide the decision-making process regarding pharmacological adjustments andthe focus of rehabilitation strategies.

Strengths of our study include the large sample size, thevariety of disorders, and the advanced disability stages ofthe patients. Other studies investigating the effectiveness ofrehabilitation have been case reports, small pretest-posttesttrials with a sample size of 19 or less, and a pilot RCTwith a sample size of 17 [6, 8–10, 12–15]. In addition, mostof the studies included subjects with PSP, whereas our studysampled across categories of atypical parkinsonism andincluded primarily vascular parkinsonism and MSA. Ourpatients were also at higher levels of disability and all wereHoehn and Yahr stages III to V, suggesting that functionalgains can be made in these patients with complex symptomsin later stages of the disease.

Limitations of the study include lack of a control groupand prescribing rehabilitation while simultaneously chang-ing medications. We are unable to distinguish what effectswere due purely to physical rehabilitation. However, oursmaller group of 10 patients who received rehabilitation onlydid show significant improvement, but larger studies shouldbe conducted to address the particular effects of rehabilita-tion alone. It is also difficult to assess whether the intensive


Table 4: Patients receiving rehabilitation only.

Measure (n)a Admission mean (SD) Discharge mean (SD) Change (95% CI)

Total FIM (10) 38.3 (13.4) 70.4 (23.4) 32.1 (22.8, 41.3)

Motor FIM (10) 21.2 (7.8) 49.6 (17.6) 28.6 (19.8, 37.2)

Cognitive FIM (10) 17.1 (6.1) 20.6 (6.1) 3.5 (1.7, 5.2)

TUG (10) 83.4 s (61.5) 31.3 s (10.1) −52.1 s (−13.7, −91.0)

TMW (9) 114.4 ft (77.6) 190.4 ft (68.4) 76.0 ft (27.4, 124.5)

Left FT (8) 61.5 (19.8) 69.3 (13.4) 7.8 (−9.7, 25.4)

Right FT (8) 67.2 (24.8) 86.8 (24.9) 19.6 (3.4, 35.7)aFIM: functional independence measure, TUG: timed up and go, TMW: 2-minute walk, BBS: berg balance scale, and FT: finger tap.

nature of the program itself had any effect on the subjects’improvement. The higher frequency of assessments and ther-apy sessions in an inpatient setting may have continued to theimprovements observed. Factors such as treatment intensity,the availability of objective data for treatment decisionsand goal setting, and the expertise and frequent commu-nication of the interdisciplinary team were not individuallyassessed. Further studies comparing specialized interdisci-plinary movement disorder rehabilitation programs such asthe program described in this study with more traditionalstandard rehabilitative care would help clarify this question.

Another limitation of our study is the lack of long-termfollowup data. It is unknown if the gains made during therehabilitation admission were sustained following discharge.One study of patients with idiopathic PD in Hoehn andYahr stages II and III who participated in an inpatient exer-cise training and muscle strengthening program sustainedimprovements in quality of life at follow-up [40]. Lastly, thegains in the total FIM used to assess clinically meaningful im-provement were extrapolated from the stroke literature [39],as this has not been derived in parkinsonian patients.

While our study demonstrated that patients with atypicalparkinsonism can benefit from an intensive inpatient rehabi-litation program, further studies are needed to look at thelong-term gains. In addition, research is needed to assessthe efficacy of inpatient rehabilitation programs on atypicalparkinsonism patients with earlier stages of disease and theireffect on the progression of their disorders.

References

[1] C. W. Chadwick and M. J. Aminoff, “Clinical differentiationof parkinsonian syndromes: prognostic and therapeutic rele-vance,” American Journal of Medicine, vol. 117, no. 6, pp. 412–419, 2004.

[2] I. Litvan, K. P. Bhatia, D. J. Burn et al., “Movement disor-ders society scientific issues committee report: SIC task force-appraisal of clinical diagnostic criteria for Parkinsonian dis-orders,” Movement Disorders, vol. 18, no. 5, pp. 467–486, 2003.

[3] I. G. McKeith, D. W. Dickson, J. Lowe et al., “Diagnosis andmanagement of dementia with Lewy bodies: third report of theDLB consortium,” Neurology, vol. 65, no. 12, pp. 1863–1872,2005.

[4] C. Colosimo and F. R. Pezzella, “The symptomatic treatmentof multiple system atrophy,” European Journal of Neurology,vol. 9, no. 3, pp. 195–199, 2002.

[5] A. E. Lang, “Treatment of progressive supranuclear palsy and-corticobasal degeneration,” Movement Disorders, vol. 20, sup-plement 12, pp. S83–S91, 2005.

[6] G. Kwakkel, C. J. T. de Goede, and E. E. H. van Wegen, “Impactof physical therapy for Parkinson’s disease: a critical reviewof the literature,” Parkinsonism and Related Disorders, vol. 13,supplement 3, pp. S478–S487, 2007.

[7] C. J. T. de Goede, S. H. J. Keus, G. Kwakkel, and R. C. Wage-naar, “The effects of physical therapy in Parkinson’s disease: aresearch synthesis,” Archives of Physical Medicine and Rehabil-itation, vol. 82, no. 4, pp. 509–515, 2001.

[8] K. L. Izzo, P. DiLorenzo, and A. Roth, “Rehabilitation in pro-gressive supranuclear palsy: case report,” Archives of PhysicalMedicine and Rehabilitation, vol. 67, no. 7, pp. 473–476, 1986.

[9] J. Sosner, G. C. Wall, and J. Sznajder, “Progressive supranuclearpalsy: clinical presentation and rehabilitation of two patients,”Archives of Physical Medicine and Rehabilitation, vol. 74, no. 5,pp. 537–539, 1993.

[10] M. Suteerawattananon, B. MacNeill, and E. J. Protas, “Sup-ported treadmill training for gait and balance in a patient withprogressive supranuclear palsy,” Physical Therapy, vol. 82, no.5, pp. 485–495, 2002.

[11] C. Zampieri and R. P. Di Fabio, “Balance and eye movementtraining to improve gait in people with progressive supranu-clear palsy: quasi-randomized clinical trial,” Physical Therapy,vol. 88, no. 12, pp. 1460–1473, 2008.

[12] C. Zampieri and R. P. Di Fabio, “Improvement of gaze controlafter balance and eye movement training in patients with pro-gressive supranuclear palsy: a quasi-randomized controlledtrial,” Archives of Physical Medicine and Rehabilitation, vol. 90,no. 2, pp. 263–270, 2009.

[13] T. M. Steffen, B. F. Boeve, L. A. Mollinger-Riemann, and C. M.Petersen, “Long-term locomotor training for gait and balancein a patient with mixed progressive supranuclear palsy andcorticobasal degeneration,” Physical Therapy, vol. 87, no. 8, pp.1078–1087, 2007.

[14] M. Landers, M. Adams, K. Acosta, and A. Fox, “Challenge-oriented gait and balance training in sporadic olivopontocere-bellar atrophy: a case study,” Journal of Neurologic PhysicalTherapy, vol. 33, no. 3, pp. 160–168, 2009.

[15] F. M. Wedge, “The impact of resistance training on balanceand functional ability of a patient with multiple systematrophy,” Journal of Geriatric Physical Therapy, vol. 31, no. 2,pp. 79–83, 2008.

[16] K. Kawahira, T. Noma, J. Iiyama, S. Etoh, A. Ogata, and M.Shimodozono, “Improvements in limb kinetic apraxia by re-petition of a newly designed facilitation exercise in a patientwith corticobasal degeneration,” International Journal of Reha-bilitation Research, vol. 32, no. 2, pp. 178–183, 2009.


[17] S. Jain, J. Dawson, N. P. Quinn, and E. D. Playford, “Occupa-tional therapy in multiple system atrophy: a pilot randomizedcontrolled trial,” Movement Disorders, vol. 19, no. 11, pp.1360–1364, 2004.

[18] T. Ellis, D. I. Katz, D. K. White, T. J. DePiero, A. D. Hohler,and M. Saint-Hilaire, “Effectiveness of an inpatient multidis-ciplinary rehabilitation program for people with Parkinsondisease,” Physical Therapy, vol. 88, no. 7, pp. 812–819, 2008.

[19] K. J. Ottenbacher, Y. Hsu, C. V. Granger, and R. C. Fiedler,“The reliability of the functional independence measure:a quantitative review,” Archives of Physical Medicine andRehabilitation, vol. 77, no. 12, pp. 1226–1232, 1996.

[20] K. E. Light, A. Behrman, M. Thigpen, and W. Triggs, “The 2-minute walk test: a tool for evaluating walking endurance inclients with Parkinson’s disease,” Journal of Neurologic PhysicalTherapy, vol. 21, no. 4, pp. 136–139, 1997.

[21] S. Morris, M. E. Morris, and R. Iansek, “Reliability of measure-ments obtained with the timed “Up, & Go” test in people withParkinson disease,” Physical Therapy, vol. 81, no. 2, pp. 810–818, 2001.

[22] M. Conradsson, L. Lundin-Olsson, N. Lindelof et al., “Bergbalance scale: intrarater test-retest reliability among olderpeople dependent in activities of daily living and living in resi-dential care facilities,” Physical Therapy, vol. 87, no. 9, pp.1155–1163, 2007.

[23] A. A. Qutubuddin, P. O. Pegg, D. X. Cifu, R. Brown, S.McNamee, and W. Carne, “Validating the Berg balance scalefor patients with Parkinson’s disease: a key to rehabilitationevaluation,” Archives of Physical Medicine and Rehabilitation,vol. 86, no. 4, pp. 789–792, 2005.

[24] T. J. Stevenson, “Detecting change in patients with stroke usingthe Berg balance scale,” Australian Journal of Physiotherapy,vol. 47, no. 1, pp. 20–38, 2001.

[25] T. Steffen and M. Seney, “Test-retest reliability and minimaldetectable change on balance and ambulation tests, the 36-item short-form health survey, and the unified Parkinsondisease rating scale in people with parkinsonism,” PhysicalTherapy, vol. 88, no. 6, pp. 733–746, 2008.

[26] I. Shimoyama, T. Ninchoji, and K. Uemura, “The finger-tapping test. A quantitative analysis,” Archives of Neurology,vol. 47, no. 6, pp. 681–684, 1990.




[30] M. H. Thaut, G. C. McIntosh, R. R. Rice, R. A. Miller, J. Rath-bun, and J. M. Brault, “Rhythmic auditory stimulation in gaittraining for Parkinson’s disease patients,” Movement Disorders,vol. 11, no. 2, pp. 193–200, 1996.

[31] M. A. Hirsch, T. Toole, C. G. Maitland, and R. A. Rider,“The effects of balance training and high-intensity resistancetraining on persons with idiopathic Parkinson’s disease,” Ar-chives of Physical Medicine and Rehabilitation, vol. 84, no. 8,pp. 1109–1117, 2003.

[32] T. Ellis, C. J. de Goede, R. G. Feldman, E. C. Wolters, G. Kwak-kel, and R. C. Wagenaar, “Efficacy of a physical therapy pro-gram in patients with Parkinson’s disease: a randomized con-trolled trial,” Archives of Physical Medicine and Rehabilitation,vol. 86, no. 4, pp. 626–632, 2005.

[33] V. Muller, B. Mohr, R. Rosin, F. Pulvermuller, F. Muller, andN. Birbaumer, “Short-term effects of behavioral treatmenton movement initiation and postural control in Parkinson’sdisease: a controlled clinical study,” Movement Disorders, vol.12, no. 3, pp. 306–314, 1997.

[34] L. E. Dibble, T. F. Hale, R. L. Marcus, J. Droge, J. P. Gerber,and P. C. LaStayo, “High-intensity resistance training amplifiesmuscle hypertrophy and functional gains in persons withParkinson’s disease,” Movement Disorders, vol. 21, no. 9, pp.1444–1452, 2006.


[36] C. L. Comella, G. T. Stebbins, N. Brown-Toms, and C. G.Goetz, “Physical therapy and Parkinson’s disease: a controlledclinical trial,” Neurology, vol. 44, no. 3, pp. 376–378, 1994.

[37] S. Sapir, L. O. Ramig, P. Hoyt, S. Countryman, C. O’Brien,and M. Hoehn, “Speech loudness and quality 12 months afterintensive voice treatment (LSVT) for Parkinson’s disease:a comparison with an alternative speech treatment,” FoliaPhoniatrica et Logopaedica, vol. 54, no. 6, pp. 296–303, 2002.

[38] L. O. Ramig, C. Fox, and S. Sapir, “Parkinson’s disease: speechand voice disorders and their treatment with the Lee Silvermanvoice treatment,” Seminars in Speech and Language, vol. 25, no.2, pp. 169–180, 2004.

[39] M. Beninato, K. M. Gill-Body, S. Salles, P. C. Stark, R. M.Black-Schaffer, and J. Stein, “Determination of the minimalclinically important difference in the FIM instrument in pa-tients with stroke,” Archives of Physical Medicine and Rehabili-tation, vol. 87, no. 1, pp. 32–39, 2006.

[40] M. E. Morris, R. Iansek, and B. Kirkwood, “A randomized con-trolled trial of movement strategies compared with exercise forpeople with Parkinson’s disease,” Movement Disorders, vol. 24,no. 1, pp. 64–71, 2009.


Review Article

A Review of Dual-Task Walking Deficits in People withParkinson’s Disease: Motor and Cognitive Contributions,Mechanisms, and Clinical Implications

Valerie E. Kelly, Alexis J. Eusterbrock, and Anne Shumway-Cook

Department of Rehabilitation Medicine, University of Washington, 1959 NE Pacific Street, P.O. Box 356490,Seattle, WA 98195-6490, USA

Correspondence should be addressed to Valerie E. Kelly, [email protected]

Received 1 June 2011; Revised 29 August 2011; Accepted 4 September 2011


Copyright © 2012 Valerie E. Kelly et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Gait impairments in Parkinson’s disease (PD) are exacerbated under dual-task conditions requiring the simultaneous performanceof cognitive or motor tasks. Dual-task walking deficits impact functional mobility, which often requires walking while performingconcurrent tasks such as talking or carrying an object. The consequences of gait impairments in PD are significant and includeincreased disability, increased fall risk, and reduced quality of life. However, effective therapeutic interventions for dual-taskwalking deficits are limited. The goals of this narrative review are to describe dual-task walking deficits in people with PD, todiscuss motor and cognitive factors that may contribute to these deficits, to review potential mechanisms underlying dual-taskdeficits, and to discuss the effect of therapeutic interventions on dual-task walking deficits in persons with PD.

1. Introduction

Gait impairments and walking limitations are commonamong people with Parkinson’s disease (PD). While gaitabnormalities are not pronounced in the early stages ofPD, their prevalence and severity increase with diseaseprogression. Within 3 years of diagnosis, over 85% of peoplewith clinically probable PD develop gait problems [1].The potential consequences of gait impairments in PD aresignificant and include increased disability [2, 3], increasedrisk for falls, and reduced quality of life. Falls are commonamong people with PD and can result in fear of falling,injury, and hospitalization [4–10]. The estimated prevalenceof falls in PD ranges from 40 to 90% and increases with theduration of follow-up [4, 5, 11–16]. It is estimated that 45–50% of falls in this population occur when walking [5, 17],with balance and walking deficits commonly identified asrisk factors for falls [5, 10–12, 14, 18, 19]. Reduced qualityof life is also associated with balance and gait abnormalitiesin PD, including festination and freezing of gait [2, 20–24]. In fact, people with PD consider mobility and walkinglimitations to be among the worst aspects of the disease [25].

Mobility in daily life frequently requires walking whileperforming simultaneous cognitive or motor tasks, suchas talking with a friend or carrying a cup of coffee. Gaitimpairments in people with PD are exacerbated undersuch dual-task conditions. In recent years, dual-task walkingresearch has expanded rapidly. The association of gaitimpairments with adverse consequences like increased fallrisk has motivated research into clinical strategies to assessand treat dual-task walking deficits in PD. Several recentreview papers have been published on dual-task posture andgait deficits among older adults and in a general neurologicpopulation [26–29], but none have focused specifically onpeople with PD. While people with PD demonstrate dual-task deficits in a variety of movements, including posturalcontrol tasks [30, 31], upper extremity movements [32, 33],and speech [34], the focus of this paper is dual-task walking.The goals of this review are to describe dual-task walkingdeficits in people with PD, to discuss motor and cognitivefactors that may contribute to these deficits, to reviewpotential mechanisms underlying dual-task interference, andto discuss the effect of therapeutic interventions on dual-taskwalking deficits in people with PD.


2. Dual-Task Walking Deficits in PD

Single-task gait impairments in PD include reduced speedand stride length and increased double limb support timeand stride-to-stride variability [35–38]. With progressionof PD, gait abnormalities worsen, and festination, freez-ing, and dystonic or dyskinetic gait patterns can emerge[39]. Gait impairments in PD are exacerbated under dual-task conditions, with further reductions in gait speed andstride length [40–46], decreased symmetry and coordinationbetween left and right steps [47, 48], and increased stride-to-stride variability [45, 49, 50]. This section will reviewreported dual-task walking deficits in people with PD andwill consider factors that influence the magnitude of thesedeficits.

2.1. Individual, Task, and Environment Framework. Table 1summarizes dual-task walking studies in people with PD,including relevant individual, task, and environmental char-acteristics of each study. Comparing dual-task walkingdeficits across studies is challenging because of variationsin methodology. In Table 1, decrements in walking underdual-task conditions are expressed as a percentage of single-task performance, commonly referred to as the dual-taskcost (DTC = [dual-task – single-task]/single-task ∗ 100)[51, 52]. The DTC allows a more direct comparison of dual-task deficits across studies and provides a way to assessthe relative effects of individual, task, and environmentalfactors. For example, a study by Plotnik and colleaguesmeasured gait speed DTCs of 17% in people with moderatePD, on medication, when walking approximately 80 m andperforming serial-3 subtractions [45]. Lord and colleaguesmeasured gait speed DTCs of 32% in people with moderatePD, off medication, when walking approximately 6.5 m intheir home while carrying a tray and counting auditorytones [43]. Dual-task walking deficits can be comparedusing the DTC even though these studies varied in termsof the participants’ medication status, the concurrent tasksused, and the environment where walking occurred. Becausemultiple factors differed between studies, it is not clearwhether the greater DTCs reported by Lord and colleaguesare due to off-medication status, more challenging concur-rent tasks, or a more complex home environment. Whenassessing dual-task deficits in PD, it is important to considerindividual characteristics such as the severity of motor andcognitive impairments, the complexity of both walking andconcurrent tasks, and the overall challenge presented by theenvironment.

2.2. Individual Factors. Studies of dual-task walking in PDvary substantially with respect to participant characteris-tics. Dual-task walking deficits increase with age amonghealthy adults [29, 60, 61], but people with PD consistentlydemonstrate greater dual-task walking deficits than healthy,age-matched individuals [42, 44, 50, 54, 59]. For example,O’Shea and colleagues found that people with PD hadgreater dual-task declines in gait speed than healthy older

adults, with gait speed DTCs of −18% to −19% in the PDgroup compared to −7% in the control group [44]. Mostresearch has examined people with mild-to-moderate diseaseseverity, as measured by the Unified Parkinson DiseaseRating Scale (UPDRS) and Hoehn and Yahr scores, althoughdisease severity is associated with dual-task walking deficits[43, 57]. The majority of studies examined the impactof concurrent task performance during the on-medicationstate, though a small number of studies examined dual-taskwalking in people with PD in the off-medication state only[43, 59]. Studies that examined the effects of medicationdemonstrated improvements in dual-task walking perfor-mance on-medication compared to off-medication [53, 57].Some studies specifically examined individuals with PD andfreezing [53, 55, 59], motor response fluctuations [45], or ahistory of falls [62]. For example, research comparing peoplewith PD and freezing to those without freezing demonstratedincreased dual-task walking deficits when walking forwards,turning, and walking backwards [53, 55, 59].

2.3. Task Factors. Dual-task studies in PD also vary interms of walking and concurrent task characteristics. Mostexamined walking on a level surface at a self-selectedspeed, but some included more complex walking tasks. Forexample, some walking tasks involved sit-to-stand transfersand/or turning [43, 46, 53, 54, 57–59], and one studyexamined backwards walking [55]. Concurrent tasks variedin terms of type (cognitive or motor), domain, and difficulty.Concurrent cognitive tasks included mental tracking, such asattentional tasks [43, 50, 59] or arithmetic calculations [41,42, 44, 45, 47–50, 55, 56], verbal fluency or conversationaltasks [42, 53, 54], and memory tasks [46, 50]. Concurrentmotor tasks were used less commonly and included carryingobjects [40, 43, 46, 57, 58] or manipulating objects [42, 44].It is not clear whether motor or cognitive tasks have a greaterimpact on walking in people with PD. One study foundsimilar impacts of cognitive and motor tasks [44], whileother studies showed a greater impact of cognitive tasks[42, 46]. However, the tasks incorporated differed in terms ofboth type and complexity, limiting the ability to make directcomparisons. Studies that controlled task domain and variedtask difficulty suggest that more complex tasks have a greatereffect on walking in PD [40, 54, 56].

Typically, no specific instructions are provided regardingwhich task to prioritize during dual-task conditions. In mostcases, participants were either instructed to focus on bothtasks or instructions were not specified. However, moststudies quantified dual-task changes in walking only and didnot measure concurrent task performance, making it difficultto determine if there were between-task trade-offs. DTCsprovide a means to assess trade-offs between walking andconcurrent task performance [63]. In studies that examineddual-task changes in both walking and the concurrent task,most showed declines in both [44, 50]. Only one studydemonstrated concurrent task improvements and walkingdeclines under dual-task conditions [42], consistent withtrade-offs between tasks and prioritization of the concurrenttask over walking.


Ta

ble

1:Su

mm

ary

ofst

udi

esex

amin

ing

dual

-tas

kw

alki

ng

inpe

ople

wit

hP

D.R

elev

ant

indi

vidu

al,t

ask,

and

envi

ron

men

tala

spec

tsof

each

stu

dyar

ein

clu

ded

.Du

al-t

ask

cost

sfo

rw

alki

ng

and

the

con

curr

ent

task

are

incl

ude

dw

her

eth

eyco

uld

beca

lcu

late

d.

Indi

vidu

alC

har

acte

rist

ics

Task

and

Env

iron

men

talC

har

acte

rist

ics

Res

ult

sSt

udy

(PD

sam

ple)

Age

(yrs

)D

isea

sese

veri

tyC

og.

Wal

kin

gta

sk&

envi

ron

men

tC

oncu

rren

tta

skIn

stru

ct.

Wal

kin

gD

TC

Con

curr

ent

task

DT

CB

ond

and

Mor

ris

,200

0[4

0](n=

12)

65(1

0)W

ebst

ersc

ore:

13(5

)

STM

S:30

(4)

Wal

k10

m(c

omfo

rtab

lepa

ce)

Mot

or:(

1)ca

rry

tray

;(2

)ca

rry

tray

wit

hem

pty

glas

ses

No

prio

riti

zati

on

Spee

d:(1

)−2

%;(

2)−1

1%St

ride

len

gth

:(1

)+

1%;(

2)−1

%D

LS:(

1)−2

%;(

2)−2

%

—

Bro

wn

etal

.,20

09[4

1](n=

10)

67(7

)

UP

DR

S:28

(2)

H&

Y:2.

3(0

.3)

MM

SE:

≥26

Wal

k10

m(s

elf-

sele

cted

pace

,u

nob

stru

cted

wal

kway

)

Cog

nit

ive:

seri

al-3

subt

ract

ion

(dat

apr

esen

ted

for

no

mu

sic

tria

ls)

Non

epr

ovid

ed

Spee

d:−2

0%St

ride

len

gth

:−6%

DLS

:−12

%—

Cam

icio

liet

al.,

1998

[53]

(n=

10)

67(9

)O

ff-m

edO

n-m

ed

UP

DR

S:15

(4)

UP

DR

S:12

(6)

MM

SE:

27(1

)

Wal

k4.

6m

,tu

rn18

0◦,w

alk

4.6

m(s

elf-

sele

cted

pace

)

Cog

nit

ive:

verb

alfl

uen

cy(r

ecit

em

ale

nam

es)

Not

spec

ified

Step

s:−4

0%T

ime:−5

0%St

eps:−1

6%T

ime:−3

1%

— —

Cam

pbel

let

al.,

2003

[54]

(n=

9)

74(7

)H

&Y:

2.8

(0.8

)—

Tim

edU

pan

dG

o(3

m;

com

fort

able

pace

)

Cog

nit

ive:

(1)

rep

eat

phra

se;(

2)re

pea

tda

ysof

the

wee

kba

ckw

ards

Not

spec

ified

Step

s:(1

)+

1%;(

2)−1

3%T

ime:

(1)−1

%;(

2)−3

1%—

Gal

letl

yan

dB

rau

er,2

005

[42]

(n=

16)

65(1

0)U

PD

RS:

14(6

)M

MSE

:28

(3)

Wal

k10

m(c

omfo

rtab

lepa

ce)

Mot

or:(

1)bu

tton

pres

sC

ogn

itiv

e:(2

)se

rial

-3su

btra

ctio

n;(

3)ve

rbal

flu

ency

“Con

cen

trat

eon

both

task

s”

Spee

d:(1

)−7

%;

(2)−2

1%;(

3)−2

1%St

ride

len

gth

:(1)−4

%;

(2)−1

6%;(

3)−1

6%

Mot

or:

(1)

+16

%C

ogn

itiv

e:(2

)+

31%

;(3

)+

73%

Hac

kney

and

Ear

har

t,20

09[5

5](n=

78)

65(1

0)U

PD

RS:

28(9

)—

Wal

k5

m(c

omfo

rtab

lepa

ce)

both

forw

ard

and

back

war

d

Cog

nit

ive:

men

tal

arit

hm

etic

(ser

ial-

3,4,

and

6su

btra

ctio

n)

Not

spec

ified

∗ Spe

ed:−

33%

∗ Str

ide

len

gth

:−14

%—

Hau

sdor

ffet

al.,

2003

[49]

(n=

10)

Ran

ge:

52–8

2U

PD

RS:

14H

&Y:

3.1

MM

SE:

27W

alk

20m

(nor

mal

pace

;lev

elgr

oun

d)C

ogn

itiv

e:se

rial

-7su

btra

ctio

n

“Wal

kw

hile

perf

orm

ing

subt

ract

ion

s”

Stri

deti

me:−1

0%St

ride

tim

eva

riab

ility

:−15

4%—

LaPo

inte

etal

.,20

10[5

6](n=

25)

67H

&Y:

2.4

(7)

DR

S-2:

136

(7)

Wal

k4.

3m

Cog

nit

ive:

(1)

cou

nt

by1’

s;(2

)se

rial

-3su

btra

ctio

n;(

3)re

cite

alph

a-n

um

eric

sequ

ence

Not

spec

ified

∗∗Sp

eed:

(1)−2

%;(

2)−1

2%;(

3)−1

9%∗∗

Stri

dele

ngt

h:(

1)−1

%;

(2)−6

%;(

3)−1

0%D

LS:(

1)−4

%;(

2)−9

%;

(3)−1

0%

—


Ta

ble

1:C

onti

nu

ed.

Indi

vidu

alC

har

acte

rist

ics

Task

and

Env

iron

men

talC

har

acte

rist

ics

Res

ult

sSt

udy

(PD

sam

ple)

Age

(yrs

)D

isea

sese

veri

tyC

og.

Wal

kin

gta

sk&

envi

ron

men

tC

oncu

rren

tta

skIn

stru

ct.

Wal

kin

gD

TC

Con

curr

ent

task

DT

C

Lord

etal

.,20

10[4

3](n=

29)

71(7

)O

ff-m

edU

PD

RS:

39(1

5)

MM

SE:

27(3

)

Stan

dfr

oma

chai

r,w

alk

5–11

m(p

refe

rred

spee

d;ex

amin

edin

hom

e&

dist

ance

vari

edby

hom

e)

Mot

or:(

1)ca

rry

tray

wit

h2

beak

ers

ofw

ater

Cog

nit

ive:

(2)

cou

nt

audi

tory

ton

es(3

)M

otor

+C

ogn

itiv

e

“Con

cen

trat

eeq

ual

lyon

wal

kin

gan

dta

sk(s

)”

Spee

d:(1

)−2

4%;

(2)−1

4%;(

3)−3

2%—

Lord

etal

.,20

11[5

7](n=

50)

69(7

)O

ff-m

edO

n-m

ed

UP

DR

S:35

(9)

UP

DR

S:23

(9)

MM

SE:

28(2

)

Wal

k6

m,t

urn

180◦

,w

alk

6m

(exa

min

edin

hom

e)

Mot

or:c

arry

ing

atr

ayw

ith

2cu

psof

wat

er

“Con

cen

trat

eon

task

asa

wh

ole”

Spee

d:−1

3%St

ride

tim

eV

aria

bilit

y:−8

%Sp

eed:−1

1%St

ride

tim

eV

aria

bilit

y:−2

1%

—

O’S

hea

etal

.,20

02[4

4](n=

15)

68(7

)M

odifi

edW

ebst

erSc

ale:

12(6

)

STM

S:30

(3)

Wal

k10

m(p

refe

rred

pace

)

Mot

or:(

1)co

intr

ansf

erC

ogn

itiv

e:(2

)se

rial

-3su

btra

ctio

n

Not

spec

ified

Spee

d:(1

)−1

8%;(

2)−1

9%St

ride

len

gth

:(1)−1

4%;

(2)−1

2%D

LS:(

1)−3

%;(

2)−6

%

Mot

or:(

1)−1

7.4%

Cog

nit

ive:

(2)−5

%

Plo

tnik

etal

.,20

09[4

8](n=

21)

72(7

)U

PD

RS:

20(8

)M

MSE

:28

(1)

Wal

k2

min

ina

leve

l,25

mco

rrid

or(c

omfo

rtab

lepa

ce)

Cog

nit

ive:

seri

al-7

subt

ract

ion

No

prio

riti

zati

onP

has

eco

ordi

nat

ion

inde

x:−4

7%—

Plo

tnik

etal

.,20

11[4

5](n=

30)

66(7

)

UP

DR

S:35

(10)

H&

Y:2.

1(0

.6)

MM

SE:

29(1

)

Wal

k∼8

0m

ina

leve

l,∼2

0m

corr

idor

(com

fort

able

pace

)

Cog

nit

ive:

(1)

seri

al-3

subt

ract

ion

;(2

)se

rial

-7su

btra

ctio

n

Not

spec

ified

Spee

d:(1

)−1

7%;(

2)−2

3%St

ride

len

gth

:(1)

−11%

;(2)−1

5%St

ride

tim

eva

riab

ility

:(1)−3

9%;

(2)−5

1%

—

Roc

hes

ter

etal

.,20

04[4

6](n=

20)

65(8

)H

&Y:

2.7

(0.7

)M

MSE

:27

(2)

Stan

dfr

oma

chai

r,w

alk

6.6

(1.5

)m

,ret

urn

(pre

ferr

edsp

eed,

exam

ined

inh

ome

&di

stan

ceva

ried

byh

ome)

Mot

or:(

1)ca

rryi

ng

tray

wit

h2

cups

ofw

ater

;Cog

nit

ive:

(2)

auto

biog

raph

ical

mem

ory

task

;(3)

Mot

or+

Cog

nit

ive

Not

spec

ified

Spee

d:(1

)−9

%;(

2)−2

1%;

(3)−2

3%St

eple

ngt

h:(

1)−9

%;(

2)−1

6%;(

3)−2

1%

—

Roc

hes

ter

etal

.,20

08[5

8](n=

130)

67(8

)U

PD

RS:

33(1

1)M

MSE

:28

(2)

Wal

k6

m,t

urn

180◦

,w

alk

6m

(pre

ferr

edpa

ce)

Mot

or:c

arry

ing

atr

ayw

ith

2cu

psof

wat

er

“Con

cen

trat

eeq

ual

lyon

all

task

s”Sp

eed:−1

3%—


Ta

ble

1:C

onti

nu

ed.

Indi

vidu

alC

har

acte

rist

ics

Task

and

Env

iron

men

talC

har

acte

rist

ics

Res

ult

sSt

udy

(PD

sam

ple)

Age

(yrs

)D

isea

sese

veri

tyC

og.

Wal

kin

gta

sk&

envi

ron

men

tC

oncu

rren

tta

skIn

stru

ct.

Wal

kin

gD

TC

Con

curr

ent

task

DT

CSp

ildoo

ren

etal

.,20

10[5

9]

Free

zers

(n=

14)

69(7

)O

ff-m

ed

UP

DR

S:38

(14)

H&

Y:2.

5(0

.5)

MM

SE:

28(1

)

Wal

k5

m:(

1)st

raig

ht;

(2)

turn

180◦

;(3)

turn

360◦

Cog

nit

ive:

colo

rid

enti

fica

tion

(au

dito

ryat

ten

tion

alta

sk)

No

prio

riti

zati

on

Step

s:(1

)−2

5%;(

2)−1

6%;(

3)−1

3%T

ime:

(1)−2

3%;(

2)−1

3%;(

3)−1

0%

—

Non

-fre

ezer

s(n=

14)

67(7

)O

ff-m

ed

UP

DR

S:34

(10)

H&

Y:2.

4(0

.3)

MM

SE:

29(1

)

Step

s:(1

)−7

%;(

2)+

1%;

(3)−2

%T

ime:

(1)−9

%;(

2)−1

%;

(3)−3

%

—

Yoge

vet

al.,

2005

[50]

(n=

30)

71(8

)U

PD

RS:

18(8

)H

&Y:

2.3

(0.4

)

MM

SE:

28(2

)

Wal

k2

min

ina

leve

l,25

mco

rrid

or(c

omfo

rtab

lepa

ce)

Cog

nit

ive:

(1)

liste

nto

ata

pe

&an

swer

ques

tion

s;(2

)ab

ove

task

+ph

onem

em

onit

orin

g;(3

)se

rial

-7su

btra

ctio

n

No

prio

riti

zati

on

Spee

d:(1

)−1

0%;(

2)−1

3%;(

3)−1

9%St

ride

tim

eva

riab

ility

:(1)−1

%;

(2)−6

%;(

3)−2

7%

Cog

nit

ive:

(1)−4

2%

Yoge

vet

al.,

2007

[47]

(n=

21)

72(7

)U

PD

RS:

20(8

)H

&Y:

ran

ge2-

3

MM

SE:

28(1

)

Wal

k2

min

ina

leve

l,25

mco

rrid

or(c

omfo

rtab

lepa

ce)

Cog

nit

ive:

seri

al-7

subt

ract

ion

No

prio

riti

zati

onG

ait

asym

met

ry:−

43%

—

∗:

data

colla

psed

acro

ssfo

rwar

d&

back

war

dw

alki

ng;∗∗

:da

taco

llaps

edac

ross

PD

&co

ntr

olgr

oups

;C

og.:

cogn

itiv

est

atu

s;D

LS:

dou

ble

limb

supp

ort;

DR

S-2:

Dem

enti

aR

atin

gSc

ale-

2;D

TC

:du

al-t

ask

cost

;H

&Y:

Hoe

hn

&Ya

hr

stag

e;In

stru

ct.:

inst

ruct

ion

spr

ovid

eddu

rin

gdu

al-t

ask

con

diti

ons;

MM

SE:M

ini-

Men

talS

tate

Exa

min

atio

n;M

oCA

:Mon

trea

lCog

nit

ive

Ass

essm

ent;

STM

S:Sh

ortT

esto

fMen

talS

tatu

s;St

eps:

nu

mbe

rof

step

sre

quir

edto

com

plet

ew

alki

ng

task

;Tim

e:ti

me

requ

ired

toco

mpl

ete

wal

kin

gta

sk;U

PD

RS:

Un

ified

Park

inso

nD

isea

seR

atin

gSc

ale,

mot

orex

amin

atio

n.


2.4. Environmental Factors. Studies that systematically ma-nipulate environmental factors to determine the effects ondual-task walking deficits in PD are lacking. Most researchwas conducted in a clinical or laboratory environment, butsome was conducted in participants’ homes [43, 46, 57, 58].Studies conducted in the home environment may be morerepresentative of mobility challenges in daily life.

In summary, the literature as a whole confirms thepresence of significant dual-task walking deficits amongpersons with PD, despite methodological variations in par-ticipant characteristics, task demands, and environmentalconstraints. The extent of these deficits appears to vary as afunction of individual, task, and environmental characteris-tics, but the relative contribution of each factor is not wellunderstood. Carefully controlled studies are needed to betterquantify how these factors impact dual-task walking deficitsin people with PD.

3. Motor and Cognitive Factors Contributing toDual-Task Walking Deficits

3.1. Motor Factors. It is not clear how motor and cognitivesymptoms contribute to either single-task or dual-taskwalking deficits in PD. The motor phenotype of PD isheterogeneous, with cardinal features of rigidity, tremor,and bradykinesia [64]. These symptoms, as well as primaryimpairments in locomotor control pathways [65], can con-tribute to both single- and dual-task gait abnormalities. Therelative contributions of these factors may vary with diseaseprogression. Cardinal symptoms may contribute more towalking deficits early in the disease, while primary gaitimpairments might predominate later in the disease.

Single-task walking deficits have been associated with avariety of motor symptoms in PD. For example, increasedaxial rigidity is associated with poorer performance onsingle-task measures of balance and functional mobility[66, 67]. In addition, rigidity may contribute to reducedlower extremity joint excursions and a forward flexed posturewhen walking [39]. Bradykinesia can lead to shortenedstep length and reduced gait speed during walking [39].Postural instability, another common motor symptom, maycontribute to gait impairments such as increased stride-to-stride variability and double limb support.

Several motor factors are associated with dual-task walk-ing deficits in PD. Dual-task gait speed has been associatedwith disease severity, as measured by Hoehn and Yahr stage[46] and UPDRS motor subscale scores [43]. The severityof PD motor symptoms has also been related to single- anddual-task gait variability both off and on medication [57].Dual-task walking performance in people with PD has beenassociated with performance-based measures of balance [46].Though not a specific motor symptom of PD, some [46],but not all [43, 57], studies have found associations betweenphysical fatigue and dual-task walking deficits in PD. Dual-task walking deficits in PD are also associated with primarygait deficits. Dual-task changes in speed and stride lengthwere associated with performance on single-task mobilitytests in people with PD [45]. In addition, dual-task walking

deficits were greater in people with PD and freezing of gaitcompared to those without freezing [53, 55, 59]. Althoughdual-task walking deficits have been associated with bothmotor symptom severity and primary gait impairments, therelative contribution of each to dual-task walking deficits hasnot been well quantified.

3.2. Cognitive Factors. PD is associated with a variety of cog-nitive impairments, including executive function, attention,memory, language, and visuospatial impairments [68–70],that could contribute to dual-task walking deficits. Cognitiveprofiles in PD are variable [71] and range from mild deficitsin specific cognitive domains to severe dementia affectingmultiple domains. It is estimated that 19–30% of people withearly, newly-diagnosed PD present with cognitive impair-ments [72–74], and these impairments worsen with diseaseprogression [69]. The presence of mild cognitive impairmentin people with PD is associated with development ofdementia within 4 years [75]. The prevalence of dementiain PD is estimated at 26–44% [76, 77], with over 80% ofpeople developing dementia within 20 years of diagnosis[13]. Depression can exacerbate cognitive impairments inPD [78], and the frequency of depression in PD is estimatedat 25–33% [79, 80].

Specific cognitive functions, such as set shifting, dividedor alternating attention, and response inhibition, may beparticularly relevant to dual-task walking [28]. Dual-taskwalking deficits in PD have been associated with impair-ments in executive function, set-shifting, and attention [43,45, 46]. For example, Plotnik and colleagues [45] demon-strated a relationship between set shifting, as measured bythe Trail Making Test, and dual-task changes in gait speedand step length. Dual-task changes in gait variability wererelated to executive function, including set shifting andglobal cognition [45, 50, 57]. Executive function, measuredby the Brixton test, has also been associated with gait speed[46] and gait speed DTCs [43]. Deficits in attention wereassociated with greater deficits in gait variability [57] andincreased gait speed DTCs [43]. Finally, depression has beenrelated to gait speed declines and gait variability increasesunder dual-task conditions in some studies [46, 57], thoughassociations between dual-task parameters and affect (bothdepression and anxiety) were not supported by all studies[45].

Cognitive impairments can contribute to dual-task walk-ing deficits in various ways. First, they may limit the ability tocompensate for gait impairments using cognitive strategies.People with PD are often taught conscious strategies toimprove their gait pattern, such as focusing on walking withlonger steps. The type and severity of cognitive impairmentsmay limit the ability to use such strategies to compensatefor gait abnormalities. Also, impaired executive functionmight result in the inappropriate or unsafe prioritization oftasks when walking under dual-task conditions. Bloem andcolleagues have proposed that increased fall risk in peoplewith PD may result in part from a “posture second” prior-itization strategy, in which concurrent tasks are prioritizedabove walking [81, 82]. Consistent with this idea, falls in


PD have been associated with reduced performance on avariety of cognitive measures [83, 84]. The prevalence ofcognitive impairments in PD and their associations withdual-task walking deficits suggest that they are an importantcontributing factor. Further research is needed, however,because little is known about how the domains and severityof cognitive impairments affect dual-task walking deficitsand their response to therapeutic interventions.

4. Potential Mechanisms UnderlyingDual-Task Walking Deficits

The mechanisms responsible for interference between walk-ing and concurrent cognitive or motor tasks in peoplewith PD are not clear. Because multiple factors contributeto dual-task walking deficits, it is likely that a numberof different mechanisms contribute to these deficits. Inaddition, characteristics of the concurrent task, such astype, domain, and difficulty, will impact the mechanismsand resources involved in dual-task performance. Thissection will review both nonspecific mechanisms proposedto explain dual-task interference across populations as wellas specific mechanisms that may contribute to dual-taskwalking deficits in PD.

4.1. Nonspecific Mechanisms. Two general theoretical frame-works have been proposed to explain dual-task interference.Capacity theory conceptualizes the information processingneeded for dual-task performance as a flexible but limitedresource [27, 85, 86]. Performance of any given task, likewalking, requires some portion of this capacity. When twotasks are performed concurrently, competition for limitedresources results in dual-task interference and deteriorationin performance of one or both tasks [26]. According to thistheory, information processing resources such as attentioncan be flexibly allocated between tasks, with many factorspotentially influencing resource allocation [86]. For example,differences in dual-task performance can result from indi-vidual differences in overall capacity, and intra-individualvariability in dual-task performance can arise from transientvariations in effective capacity due to factors like motivation,fatigue, or arousal [86]. Task-related factors also influenceresource allocation. For example, a recent meta-analysisdemonstrated that dual-task gait speed declines varied as afunction of the concurrent cognitive task in healthy youngand older adults and a general neurologic population [29].

A second general theory to explain dual-task interferenceis the bottleneck theory [87]. According to this theory, dual-task performance requires serial or sequential processingof the two concurrent tasks. Dual-task interference resultswhen two tasks compete for the same processing resources.In order to complete one task, processing of the secondtask is temporarily postponed, resulting in performancedecrements in the second task. Dual-task walking studiesare limited in their ability to discriminate between thesetwo theories, but these general mechanisms may informmethodological choices and subsequent interpretations.

4.2. PD-Specific Mechanisms. Several mechanisms specific toPD may also contribute to dual-task walking deficits. Thesemechanisms are not mutually exclusive, but might overlapwith one another. Consistent with the capacity theory, a firstspecific mechanism in people with PD is reduced movementautomaticity. Automaticity refers to the ability to performa skilled movement without conscious or executive controlor attention directed toward the movement [88, 89]. Thecontrol of standing and walking was previously thoughtto be automatic, but the role of cognitive and executivefunctions in postural control is increasingly appreciated[26, 28]. For example, in healthy young and older adults,simple reaction times increased when walking compared tositting, reflecting greater attentional demands for walking[90, 91]. The basal ganglia are proposed to play a role inthe automatic control of movement [65]. In people with PD,basal ganglia dysfunction may lead to reduced movementautomaticity and the need for increased reliance on cognitiveresources to control movements. During dual-task upperextremity movements, people with PD demonstrated greaterlevels of activity in premotor and prefrontal cortical areascompared to healthy individuals, as measured by functionalmagnetic resonance imaging [92]. Similarly, people with PDmay rely on greater cognitive control during walking, evenunder single-task conditions [37, 93]. If reduced movementautomaticity contributes to dual-task walking deficits inpeople with PD, rehabilitation strategies designed to improvethe automatic control of walking should improve dual-taskwalking.

A second mechanism that could contribute to dual-taskwalking deficits in PD is dopamine-mediated dysfunctionof the basal ganglia. Multiple parallel pathways through thebasal ganglia subserve different functions, including motor,cognitive, and limbic functions [94–96]. Degeneration ofdopaminergic neurons in PD appears to affect both motorand cognitive circuits within the basal ganglia. Pathologyof basal ganglia circuits that project to the dorsolateralprefrontal cortex may contribute to the executive functiondeficits that are prominent in people with PD [97, 98]. Forexample, specific deficits in set shifting, which are associatedwith dual-task walking deficits in PD [45], are thoughtto be mediated by the dorsolateral prefrontal cortex [98].Dual-task walking deficits are improved by anti-parkinsonmedications [53, 57], supporting the idea that motor andcognitive impairments are due in part to dopaminergicpathways. However, the impact of anti-parkinson medica-tions may be limited to those impairments mediated bydopamine dysfunction, and many studies demonstrate dual-task walking deficits in people with PD in the on-medicationstate.

A third mechanism that could contribute to dual-taskwalking deficits in PD is the presence of nondopaminergicpathology, which may affect both gait and cognition. It isincreasingly appreciated that the pathology of PD is notlimited to dopamine but includes other neurotransmittersystems, such as serotonin, norepinephrine (noradrenaline),or acetylcholine [71, 99, 100]. Dysfunction in multiple neu-rotransmitter systems may contribute to gait [101, 102] andcognitive impairments in PD [71]. Thus, non-dopaminergic


pathways may also contribute to dual-task walking deficitsin PD. Consistent with this idea, dual-task walking deficitspersist even when people with PD are optimally medicated[42, 44, 50, 54, 59].

In summary, research suggests a number of generaland specific mechanisms that may contribute to dual-task walking deficits in PD. These mechanisms are notmutually exclusive, and the relative contribution of each maydepend on factors like the symptom profile of the individualand the specific task combination performed under dual-task conditions. A better understanding of the mechanismsresponsible for dual-task walking deficits in PD can informnovel therapeutic approaches and enhance our ability toidentify optimal interventions.

5. Therapeutic Interventions: Impact onDual-Task Walking Deficits

The effects of various interventions on single-task walkingin PD have been well described, but there is less researchexamining the efficacy of different pharmacological, surgical,or rehabilitative therapies on dual-task walking in thispopulation. Because gait impairments in PD are exacerbatedby dual-task conditions, which are common in daily life,it is important to understand how various therapeuticinterventions affect dual-task walking.

5.1. Pharmacological Interventions. The reported effects ofanti-parkinson medications on walking in PD are variable,even under single-task conditions. Medications improveaspects of single-task walking, including gait speed and stridelength, but may not influence others, like stride-to-stridevariability [38, 103, 104], festination, and freezing of gait[39, 105, 106]. As noted above, anti-parkinson medicationsincrease speed and decrease variability during dual-taskwalking in PD [57] and even increase dual-task walkingspeed in those with freezing [53]. Neither of the abovestudies examined the effects of medication on concurrenttask performance, so it is unclear if medication-relatedimprovements were due to trade-offs between walking andthe concurrent task. Medications can have limited or adverseeffects on cognitive functions like set shifting [107] and cer-tain types of learning [108, 109] that are critical to dual-taskwalking. As a result, medications could negatively affect dual-task walking or result in dual-task walking improvements atthe expense of concurrent cognitive task performance. Thepositive effects of anti-parkinson medications on dual-taskwalking are consistent with a contribution from dopaminer-gic mechanisms, but persistent deficits in the on-medicationstate suggest that non-dopaminergic mechanisms may alsocontribute to dual-task interference.

5.2. Surgical Interventions. The reported effects of surgeryon single-task walking are inconsistent. For example, initialimprovements in postural control and gait as a result ofdeep brain stimulation are not sustained beyond 2–9 years[110]. In the short term, subthalamic nucleus stimulation

can improve single-task gait speed and stride length, partic-ularly in the off-medication condition [111, 112], but theindividual response to subthalamic nucleus stimulation inthe on-medication state is variable [113]. To date, no researchhas examined the effects of deep brain stimulation or ablativesurgeries on dual-task walking in people with PD. Thelimited research on dual-task upper extremity movements isequivocal, with one study showing no effect of subthalamicnucleus stimulation [114] and one showing a decline [115].

5.3. Rehabilitation Interventions. There is considerable re-search demonstrating training-related improvements insingle-task walking in persons with PD [116–122]. However,it is not clear whether dual-task walking deficits can beimproved with practice in PD or, alternatively, whetherclinicians should teach people with PD to avoid dual-taskconditions to improve safety [123]. A variety of rehabilitationstrategies to improve dual-task walking in PD have beenstudied, with most research focusing on external cues, cog-nitive or attentional strategies, and dual-task gait training.

External visual, auditory, or somatosensory cues improveboth single- and dual-task walking in PD [42, 124–129],even among those with de novo PD [130] or cognitiveimpairment [131]. For example, Rochester and colleaguesexamined the effects of external rhythmic cues (auditory,visual, and somatosensory) on walking in people with PD[128]. Cueing therapy was provided over nine 30-minutesessions in the home and consisted of training during single-and dual-task walking and during various functional walkingtasks. Speed and step length improved during both single-and dual-task cued walking conditions. These improvementstransferred to noncued walking and were retained at 6-week follow-up testing. The authors suggest that dual-task walking improvements were likely due to improvedwalking automaticity. Based on this research, external cueingappears to improve walking under both single- and dual-task conditions in people with PD. However, studies of cuetraining vary in terms of cueing modality, training duration,tests used for outcomes assessment, and length of follow-up. Further research is needed to determine the parametersof cue training that provide the greatest and most sustainedbenefits for dual-task walking in PD.

Cognitive or attentional strategies (e.g., focusing atten-tion on walking with long steps) can also improve walking inpeople with PD [125, 126, 132], but evidence for the efficacyof cognitive strategies to improve dual-task walking is mixed.Dual-task conditions introduce a concurrent task requiringcognitive control. As suggested by the capacity theory ofdual-task interference, the need to direct cognitive resourcesto the concurrent task may limit the ability to use consciousor unconscious cognitive control to improve walking in PD.Some studies indicate that attention can improve dual-taskwalking [125], while others find that attentional strategies arenot effective under dual-task conditions [133].

Recent intervention studies have combined dual-taskgait training with cognitive strategies to direct attentionalfocus and task prioritization. Even people with early PDreport the need to monitor and consciously correct walking


deficits [93]. However, research suggests that people with PDprioritize concurrent tasks over postural tasks under dualtask conditions, thereby decreasing safety and increasing fallrisk [82]. A number of intervention studies have examinedthe effects of dual-task training with various instructionsregarding task prioritization. Training with instructions toprioritize walking improved gait velocity and stride lengthunder both single- and dual-task conditions [125, 134],with retention at 30 minutes [134]. Dual-task training withinstructions to divide attention equally between walkingand the concurrent cognitive task also improved dual-taskgait speed and stride length, with retention at 30 minutes[135]. However, the same concurrent task was used for bothtraining and outcomes measurement in this study, so it isnot clear if these training-related improvements generalizeto other dual-task combinations. Canning and colleaguesalso examined multitask training with divided attentioninstructions [136]. In this study, the concurrent tasks usedduring training differed from those used for outcomesmeasurement. Training improved gait speed and cadence,with improvements retained at 3-week follow-up. Finally,Brauer and Morris examined the effects of dual-task trainingusing variable-priority instructions, where prioritization isshifted between walking and the concurrent task [137].Gait speed and step length improved for both the traineddual-task combinations and on novel dual-task walkingcombinations. Performance on the concurrent tasks didnot decline, indicating that dual-task walking improvementswere not due to between-task trade-offs. The authors suggestthat practice may reduce the attentional demands of walkingand increase automaticity, thus enabling individuals with PDto attend to more challenging concurrent tasks. Together,these studies suggest that dual-task gait training is aneffective intervention, but the relative impact of differentinstructional sets requires further research.

One of the limitations in the research on dual-taskwalking interventions is the lack of consistent and validatedmeasures of dual-task walking performance. Appropriateoutcome measures are necessary to determine if a personwith PD has dual-task walking deficits and if a givenintervention effectively improves these deficits. A varietyof tests, including the Stops Walking When Talking testor the Walking and Remembering Test, have been used toassess dual-task walking performance in older adults [138–144]. Few of these measures have been examined in the PDpopulation [54, 81, 145], thus the psychometric propertiesof these tests in PD are unclear. Future research is needed todetermine reliable, valid, and sensitive outcome measures toevaluate dual-task walking performance in people with PDand quantify the response to different interventions.

Research supports the efficacy of rehabilitative inter-ventions, including external cueing, cognitive strategies,and dual-task gait training, to improve dual-task walkingdeficits in PD. Emerging research is examining additionaltreatment approaches to improve dual-task walking. Forexample, treadmill training with virtual reality, designed toincorporate more complex task and environmental condi-tions, has been shown to improve both single- and dual-task walking in people with PD [146]. Future research is

needed to examine optimal treatment parameters for bothestablished and novel dual-task walking interventions, therelative efficacy of different interventions, whether dual-task walking improvements generalize to novel dual-taskcombinations, and the degree to which improvements indual-task walking are retained.

6. Summary

This paper has reviewed basic and applied research relatedto dual-task walking deficits in people with PD. Gait impair-ments under both single-task and dual-task conditions areprevalent in people with PD and are associated with seriousconsequences. The severity of dual-task walking deficitsappears to vary as a function of individual, task, andenvironmental characteristics, though the relative impactsof each factor are not well understood. Both motor andcognitive impairments have been associated with dual-taskwalking deficits in persons with PD. However, because theclinical profile of PD is heterogeneous, further research isneeded to elucidate the relative contributions of each ofthese impairments to dual-task walking deficits. A numberof general and specific mechanisms may underlie dual-taskwalking deficits in PD. The role of each is not clear, butmight depend on the dual-task combination performed.These mechanisms inform a number of therapeutic interven-tions. Rehabilitation interventions, including external cues,cognitive strategies, and dual-task gait training, appear tobe effective in reducing dual-task walking deficits in PD.However, a better understanding of the individual, task,and environmental factors that influence dual-task walkingdeficits is critical to refine existing interventions and identifynovel therapeutic approaches.

Acknowledgment

This work was supported by the National Institutes of Health,National Institute of Child Health and Human Development(K01HD052018).

References

[1] G. A. Kang, J. M. Bronstein, D. L. Masterman, M. Redelings,J. A. Crum, and B. Ritz, “Clinical characteristics in earlyParkinson’s disease in a central California population-basedstudy,” Movement Disorders, vol. 20, no. 9, pp. 1133–1142,2005.

[2] D. Muslimovic, B. Post, J. D. Speelman, B. Schmand, and R.J. de Haan, “Determinants of disability and quality of life inmild to moderate Parkinson disease,” Neurology, vol. 70, no.23, pp. 2241–2247, 2008.

[3] B. Post, M. P. Merkus, R. J. de Haan, and J. D. Speelman,“Prognostic factors for the progression of Parkinson’s disease:a systematic review,” Movement Disorders, vol. 22, no. 13, pp.1839–1851, 2007.

[4] W. C. Koller, S. Glatt, B. Vetere-Overfield, and R. Hassanein,“Falls and Parkinson’s disease,” Clinical Neuropharmacology,vol. 12, no. 2, pp. 98–105, 1989.

[5] P. Gray and K. Hildebrand, “Fall risk factors in Parkinson’sdisease,” The Journal of Neuroscience Nursing, vol. 32, no. 4,pp. 222–228, 2000.


[6] C. L. Wielinski, C. Erickson-Davis, R. Wichmann, M. Walde-Douglas, and S. A. Parashos, “Falls and injuries resultingfrom falls among patients with Parkinson’s disease and otherParkinsonian syndromes,” Movement Disorders, vol. 20, no.4, pp. 410–415, 2005.

[7] H. Woodford and R. Walker, “Emergency hospital admis-sions in idiopathic’s Parkinson’s disease,” Movement Disor-ders, vol. 20, no. 9, pp. 1104–1108, 2005.

[8] R. W. Genever, T. W. Downes, and P. Medcalf, “Fracturerates in Parkinson’s disease compared with age- and gender-matched controls: a retrospective cohort study,” Age andAgeing, vol. 34, no. 1, pp. 21–24, 2005.

[9] L. J. Melton, C. L. Leibson, S. J. Achenbach et al., “Fracturerisk after the diagnosis of Parkinson’s disease: influence ofconcomitant dementia,” Movement Disorders, vol. 21, no. 9,pp. 1361–1367, 2006.

[10] P. Vestergaard, L. Rejnmark, and L. Mosekilde, “Fracture riskassociated with parkinsonism and anti-Parkinson drugs,”Calcified Tissue International, vol. 81, no. 3, pp. 153–161,2007.

[11] A. Ashburn, E. Stack, R. M. Pickering, and C. D. Ward,“A community-dwelling sample of people with Parkinson’sdisesase: characteristics of fallers and non-fallers,” Age andAgeing, vol. 30, no. 1, pp. 47–52, 2001.

[12] B. R. Bloem, Y. A. M. Grimbergen, M. Cramer, M. Willemsen,and A. H. Zwinderman, “Prospective assessment of falls inParkinson’s disease,” Journal of Neurology, vol. 248, no. 11,pp. 950–958, 2001.

[13] M. A. Hely, W. G. J. Reid, M. A. Adena, G. M. Halliday, and J.G. L. Morris, “The Sydney Multicenter Study of Parkinson’sdisease: the inevitability of dementia at 20 years,” MovementDisorders, vol. 23, no. 6, pp. 837–844, 2008.

[14] G. K. Kerr, C. J. Worringham, M. H. Cole, P. F. Lacherez,J. M. Wood, and P. A. Silburn, “Predictors of future falls inParkinson disease,” Neurology, vol. 75, no. 2, pp. 116–124,2010.


[16] B. H. Wood, J. A. Bilclough, A. Bowron, and R. W. Walker,“Incidence and prediction of falls in Parkinson’s disease:a prospective multidisciplinary study,” Journal of NeurologyNeurosurgery and Psychiatry, vol. 72, no. 6, pp. 721–725,2002.

[17] A. Ashburn, E. Stack, C. Ballinger, L. Fazakarley, and C.Fitton, “The circumstances of falls among people withParkinson’s disease and the use of Falls Diaries to facilitatereporting,” Disability and Rehabilitation, vol. 30, no. 16, pp.1205–1212, 2008.


[19] A. C. Dennison, J. V. Noorigian, K. M. Robinson et al.,“Falling in Parkinson disease: identifying and prioritizingrisk factors in recurrent fallers,” American Journal of PhysicalMedicine and Rehabilitation, vol. 86, no. 8, pp. 621–632, 2007.

[20] J. C. Gomez-Esteban, J. J. Zarranz, E. Lezcano et al.,“Influence of motor symptoms upon the quality of life ofpatients with Parkinson’s disease,” European Neurology, vol.57, no. 3, pp. 161–165, 2007.

[21] K. E. Lyons, R. Pahwa, A. I. Troster, and W. C. Koller,“A comparison of Parkinson’s disease symptoms and self-reported functioning and well being,” Parkinsonism andRelated Disorders, vol. 3, no. 4, pp. 207–209, 1997.

[22] O. Moore, C. Peretz, and N. Giladi, “Freezing of gait affectsquality of life of peoples with Parkinson’s disease beyond itsrelationships with mobility and gait,” Movement Disorders,vol. 22, no. 15, pp. 2192–2195, 2007.

[23] R. Shibley, H. J. Griffin, N. P. Quinn, and M. Jahanshahi,“Quality of life in Parkinson’s disease: the relative importanceof the symptoms,” Movement Disorders, vol. 23, no. 10, pp.1428–1434, 2008.

[24] S. E. Soh, M. E. Morris, and J. L. McGinley, “Determinantsof health-related quality of life in Parkinson’s disease: asystematic review,” Parkinsonism and Related Disorders, vol.19, no. 1, pp. 1–9, 2011.

[25] M. Schenkman, T. M. Cutson, C. W. Zhu, and K. Whetten-Goldstein, “A longitudinal evaluation of patients’ perceptionsof Parkinson’s disease,” Gerontologist, vol. 42, no. 6, pp. 790–798, 2002.

[26] M. Woollacott and A. Shumway-Cook, “Attention and thecontrol of posture and gait: a review of an emerging area ofresearch,” Gait and Posture, vol. 16, no. 1, pp. 1–14, 2002.

[27] E. V. Fraizer and S. Mitra, “Methodological and interpretiveissues in posture-cognition dual-tasking in upright stance,”Gait and Posture, vol. 27, no. 2, pp. 271–279, 2008.

[28] G. Yogev-Seligmann, J. M. Hausdorff, and N. Giladi, “Therole of executive function and attention in gait,” MovementDisorders, vol. 23, no. 3, pp. 329–342, 2008.

[29] E. Al-Yahya, H. Dawes, L. Smith, A. Dennis, K. Howells, andJ. Cockburn, “Cognitive motor interference while walking:a systematic review and meta-analysis,” Neuroscience andBiobehavioral Reviews, vol. 35, no. 3, pp. 715–728, 2011.

[30] C. Dromey, E. Jarvis, S. Sondrup, S. Nissen, K. B. Foreman,and L. E. Dibble, “Bidirectional interference between speechand postural stability in individuals with Parkinson’s disease,”International Journal of Speech-Language Pathology, vol. 12,no. 5, pp. 446–454, 2010.

[31] R. Marchese, M. Bove, and G. Abbruzzese, “Effect ofcognitive and motor tasks on postural stability in Parkinson’sdisease: a posturographic study,” Movement Disorders, vol.18, no. 6, pp. 652–658, 2003.

[32] S. D. Pradhan, B. R. Brewer, G. E. Carvell, P. J. Sparto, A.Delitto, and Y. Matsuoka, “Assessment of fine motor controlin individuals with Parkinson’s disease using force trackingwith a secondary cognitive task,” Journal of NeurologicPhysical Therapy, vol. 34, no. 1, pp. 32–40, 2010.

[33] E. L. Proud and M. E. Morris, “killed hand dexterity inParkinson’s disease: effects of adding a concurrent task,”Archives of Physical Medicine and Rehabilitation, vol. 91, no.5, pp. 794–799, 2010.

[34] A. K. Ho, R. Iansek, and J. L. Bradshaw, “The effect of aconcurrent task on Parkinsonian speech,” Journal of Clinicaland Experimental Neuropsychology, vol. 24, no. 1, pp. 36–47,2002.

[35] O. Blin, A. M. Ferrandez, and G. Serratrice, “Quantitativeanalysis of gait in Parkinson patients: increased variability ofstride length,” Journal of the Neurological Sciences, vol. 98, no.1, pp. 91–97, 1990.

[36] J. M. Hausdorff, M. E. Cudkowicz, R. Firtion, J. Y. Wei, and A.L. Goldberger, “Gait variability and basal ganglia disorders:stride-to-stride variations of gait cycle timing in Parkinson’sdisease and Huntington’s disease,” Movement Disorders, vol.13, no. 3, pp. 428–437, 1998.


[37] M. E. Morris, R. Iansek, T. A. Matyas, and J. J. Summers,“The pathogenesis of gait hypokinesia in Parkinson’s disease,”Brain, vol. 117, no. 5, part 5, pp. 1169–1181, 1994.

[38] J. D. O’Sullivan, C. M. Said, L. C. Dillon, M. Huffman, and A.J. Hughes, “Gait analysis in patients with Parkinson’s diseaseand motor fluctuations: influence of levodopa and com-parison with other measures of motor function,” MovementDisorders, vol. 13, no. 6, pp. 900–906, 1998.

[39] N. Giladi, “Gait disturbances in advanced stages of Parkin-son’s disease,” Advances in Neurology, vol. 86, pp. 273–278,2001.

[40] J. M. Bond and M. Morn’s, “Goal-directed secondary motortasks: their effects on gait in subjects with Parkinson disease,”Archives of Physical Medicine and Rehabilitation, vol. 81, no.1, pp. 110–116, 2000.

[41] L. A. Brown, N. de Bruin, J. B. Doan, O. Suchowersky, andB. Hu, “Novel challenges to gait in Parkinson’s disease: theeffect of concurrent music in single- and dual-task contexts,”Archives of Physical Medicine and Rehabilitation, vol. 90, no.9, pp. 1578–1583, 2009.

[42] R. Galletly and S. G. Brauer, “Does the type of concurrenttask affect preferred and cued gait in people with Parkinson’sdisease?” Australian Journal of Physiotherapy, vol. 51, no. 3,pp. 175–180, 2005.

[43] S. Lord, L. Rochester, V. Hetherington, L. M. Allcock, andD. Burn, “Executive dysfunction and attention contribute togait interference in ’off ’ state Parkinson’s Disease,” Gait andPosture, vol. 31, no. 2, pp. 169–174, 2010.

[44] S. O’Shea, M. E. Morris, and R. Iansek, “Dual task interfer-ence during gait in people with Parkinson disease: effects ofmotor versus cognitive secondary tasks,” Physical Therapy,vol. 82, no. 9, pp. 888–897, 2002.

[45] M. Plotnik, Y. Dagan, T. Gurevich, N. Giladi, and J.M. Hausdorff, “Effects of cognitive function on gait anddual tasking abilities in patients with Parkinson’s diseasesuffering from motor response fluctuations,” ExperimentalBrain Research, vol. 208, no. 2, pp. 169–179, 2011.

[46] L. Rochester, V. Hetherington, D. Jones et al., “Attending tothe task: interference effects of functional tasks on walkingin Parkinson’s disease and the roles of cognition, depression,fatigue, and balance,” Archives of Physical Medicine andRehabilitation, vol. 85, no. 10, pp. 1578–1585, 2004.

[47] G. Yogev, M. Plotnik, C. Peretz, N. Giladi, and J. M. Haus-dorff, “Gait asymmetry in patients with Parkinson’s diseaseand elderly fallers: when does the bilateral coordination ofgait require attention?” Experimental Brain Research, vol. 177,no. 3, pp. 336–346, 2007.

[48] M. Plotnik, N. Giladi, and J. M. Hausdorff, “Bilateralcoordination of gait and Parkinson’s disease: the effectsof dual tasking,” Journal of Neurology, Neurosurgery andPsychiatry, vol. 80, no. 3, pp. 347–350, 2009.

[49] J. M. Hausdorff, J. Balash, and N. Giladi, “Effects of cognitivechallenge on gait variability in patients with Parkinson’sdisease,” Journal of Geriatric Psychiatry and Neurology, vol.16, no. 1, pp. 53–58, 2003.

[50] G. Yogev, N. Giladi, C. Peretz, S. Springer, E. S. Simon,and J. M. Hausdorff, “Dual tasking, gait rhythmicity, andParkinson’s disease: which aspects of gait are attentiondemanding?” European Journal of Neuroscience, vol. 22, no.5, pp. 1248–1256, 2005.

[51] O. Bock, “Dual-task costs while walking increase in oldage for some, but not for other tasks: an experimentalstudy of healthy young and elderly persons,” Journal ofNeuroEngineering and Rehabilitation, vol. 5, article 27, 2008.

[52] M. Doumas, C. Smolders, and R. T. Krampe, “Task prioriti-zation in aging: effects of sensory information on concurrentposture and memory performance,” Experimental BrainResearch, vol. 187, no. 2, pp. 275–281, 2008.

[53] R. Camicioli, B. S. Oken, G. Sexton, J. A. Kaye, and J. G.Nutt, “Verbal fluency task affects gait in Parkinson’s diseasewith motor freezing,” Journal of Geriatric Psychiatry andNeurology, vol. 11, no. 4, pp. 181–185, 1998.

[54] C. M. Campbell, J. L. Rowse, M. A. Ciol et al., “The effectof cognitive demand on Timed Up and Go performance inolder adults with and without Parkinson disease,” NeurologyReport, vol. 27, no. 1, pp. 2–7, 2003.

[55] M. E. Hackney and G. M. Earhart, “The effects of a secondarytask on forward and backward walking in Parkinson’sdisease,” Neurorehabilitation and Neural Repair, vol. 24, no.1, pp. 97–106, 2009.

[56] L. L. Lapointe, J. A. G. Stierwalt, and C. G. Maitland,“Talking while walking: cognitive loading and injuriousfalls in Parkinson’s disease,” International Journal of Speech-Language Pathology, vol. 12, no. 5, pp. 455–459, 2010.

[57] S. Lord, K. Baker, A. Nieuwboer et al., “Gait variabilityin Parkinson’s disease: an indicator of non-dopaminergiccontributors to gait dysfunction?” Journal of Neurology, vol.258, no. 4, pp. 566–572, 2011.

[58] L. Rochester, A. Nieuwboer, K. Baker et al., “Walking speedduring single and dual tasks in Parkinson’s disease: whichcharacteristics are important?” Movement Disorders, vol. 23,no. 16, pp. 2312–2318, 2008.

[59] J. Spildooren, S. Vercruysse, K. Desloovere, W. Vanden-berghe, E. Kerckhofs, and A. Nieuwboer, “Freezing of gait inParkinson’s disease: the impact of dual-tasking and turning,”Movement Disorders, vol. 25, no. 15, pp. 2563–2570, 2010.

[60] A. W. Priest, K. B. Salamon, and J. H. Hollman, “Age-relateddifferences in dual task walking: a cross sectional study,”Journal of NeuroEngineering and Rehabilitation, vol. 5, article29, 2008.

[61] V. E. Kelly, M. A. Schrager, R. Price, L. Ferrucci, andA. Shumway-Cook, “Age-associated effects of a concurrentcognitive task on gait speed and stability during narrow-basewalking,” Journals of Gerontology, Series A, vol. 63, no. 12, pp.1329–1334, 2008.

[62] M. Plotnik, N. Giladi, Y. Dagan, and J. M. Hausdorff,“Postural instability and fall risk in Parkinson’s disease:impaired dual tasking, pacing, and bilateral coordination ofgait during the ”ON” medication state,” Experimental BrainResearch, vol. 210, no. 3-4, pp. 529–538, 2011.

[63] V. E. Kelly, A. A. Janke, and A. Shumway-Cook, “Effects ofinstructed focus and task difficulty on concurrent walkingand cognitive task performance in healthy young adults,”Experimental Brain Research, vol. 207, no. 1-2, pp. 65–73,2010.

[64] A. Samii, J. G. Nutt, and B. R. Ransom, “Parkinson’s disease,”Lancet, vol. 363, no. 9423, pp. 1783–1793, 2004.

[65] K. Takakusaki, J. Oohinata-Sugimoto, K. Saitoh, and T.Habaguchi, “Role of basal ganglia-brainstem systems in thecontrol of postural muscle tone and locomotion,” Progress inBrain Research, vol. 143, pp. 231–237, 2004.

[66] E. Franzen, C. Paquette, V. S. Gurfinkel, P. J. Cordo, J. G. Nutt,and F. B. Horak, “Reduced performance in balance, walkingand turning tasks is associated with increased neck tone inParkinson’s disease,” Experimental Neurology, vol. 219, no. 2,pp. 430–438, 2009.

[67] M. Schenkman, M. Morey, and M. Kuchibhatla, “Spinalflexibility and balance control among community-dwelling


adults with and without Parkinson’s disease,” Journals ofGerontology, Series A, vol. 55, no. 8, pp. M441–M445, 2000.

[68] D. Aarsland, K. Bronnick, C. Williams-Gray et al., “Mildcognitive impairment in Parkinson disease: a multicenterpooled analysis,” Neurology, vol. 75, no. 12, pp. 1062–1069,2010.

[69] D. Muslimovic, B. Schmand, J. D. Speelman, and R. J. deHaan, “Course of cognitive decline in Parkinson’s disease: ameta-analysis,” Journal of the International Neuropsychologi-cal Society, vol. 13, no. 6, pp. 920–932, 2007.

[70] G. S. Watson and J. B. Leverenz, “Profile of cognitiveimpairment in parkinson’s disease,” Brain Pathology, vol. 20,no. 3, pp. 640–645, 2010.

[71] A. A. Kehagia, R. A. Barker, and T. W. Robbins, “Neuropsy-chological and clinical heterogeneity of cognitive impairmentand dementia in patients with Parkinson’s disease,” TheLancet Neurology, vol. 9, no. 12, pp. 1200–1213, 2010.

[72] D. Aarsland, K. Brønnick, J. P. Larsen, O. B. Tysnes, andG. Alves, “Cognitive impairment in incident, untreatedparkinson disease: the norwegian parkwest study,” Neurology,vol. 72, no. 13, pp. 1121–1126, 2009.

[73] E. Elgh, M. Domellof, J. Linder, M. Edstrom, H. Stenlund,and L. Forsgren, “Cognitive function in early Parkinson’sdisease: a population-based study,” European Journal ofNeurology, vol. 16, no. 12, pp. 1278–1284, 2009.

[74] D. Muslimovic, B. Post, J. D. Speelman, and B. Schmand,“Cognitive profile of patients with newly diagnosed Parkin-son disease,” Neurology, vol. 65, no. 8, pp. 1239–1245, 2005.

[75] C. C. Janvin, J. P. Larsen, D. Aarsland, and K. Hugdahl,“Subtypes of mild cognitive impairment in Parkinson’sdisease: progression to dementia,” Movement Disorders, vol.21, no. 9, pp. 1343–1349, 2006.

[76] D. Aarsland, K. Andersen, J. P. Larsen, A. Lolk, and P. Kragh-Sørensen, “Prevalence and characteristics of dementia inParkinson disease: an 8-year prospective study,” Archives ofNeurology, vol. 60, no. 3, pp. 387–392, 2003.

[77] P. Hobson and J. Meara, “The detection of dementia andcognitive impairment in a community population of elderlypeople with Parkinson’s disease by use of the CAMCOGneuropsychological test,” Age and Ageing, vol. 28, no. 1, pp.39–44, 1999.

[78] S. Norman, A. I. Troster, J. A. Fields, and R. Brooks, “Effectsof depression and Parkinson’s disease on cognitive function-ing,” Journal of Neuropsychiatry and Clinical Neurosciences,vol. 14, no. 1, pp. 31–36, 2002.

[79] N. Giladi, T. A. Treves, D. Paleacu et al., “Risk factorsfor dementia, depression and psychosis in long-standingParkinson’s disease,” Journal of Neural Transmission, vol. 107,no. 1, pp. 59–71, 2000.

[80] O. Riedel, J. Klotsche, A. Spottke et al., “Frequency ofdementia, depression, and other neuropsychiatric symptomsin 1,449 outpatients with Parkinson’s disease,” Journal ofNeurology, vol. 257, no. 7, pp. 1073–1082, 2010.

[81] B. R. Bloem, V. V. Valkenburg, M. Slabbekoorn, and J. G.van Dijk, “The multiple tasks test. Strategies in Parkinson’sdisease,” Experimental Brain Research, vol. 137, no. 3-4, pp.478–486, 2001.

[82] B. R. Bloem, Y. A. M. Grimbergen, J. G. van Dijk, and M.Munneke, “The ”posture second” strategy: a review of wrongpriorities in Parkinson’s disease,” Journal of the NeurologicalSciences, vol. 248, no. 1-2, pp. 196–204, 2006.

[83] L. M. Allcock, E. N. Rowan, I. N. Steen, K. Wesnes, R. A.Kenny, and D. J. Burn, “Impaired attention predicts falling

in Parkinson’s disease,” Parkinsonism and Related Disorders,vol. 15, no. 2, pp. 110–115, 2009.

[84] R. Camicioli and S. R. Majumdar, “Relationship betweenmild cognitive impairment and falls in older people withand without Parkinson’s disease: 1-Year Prospective CohortStudy,” Gait and Posture, vol. 32, no. 1, pp. 87–91, 2010.

[85] M. Tombu and P. Jolicoæur, “A central capacity sharingmodel of dual-task performance,” Journal of ExperimentalPsychology: Human Perception and Performance, vol. 29, no.1, pp. 3–18, 2003.

[86] B. Abernethy, “Dual-task methodology and motor skillsresearch: some applications and methodological constraints,”Journal of Human Movement Studies, vol. 14, no. 3, pp. 101–132, 1988.

[87] E. Ruthruff, H. E. Pashler, and A. Klaassen, “Processing bot-tlenecks in dual-task performance: structural limitation orstrategic postponement?” Psychonomic Bulletin and Review,vol. 8, no. 1, pp. 73–80, 2001.

[88] R. A. Poldrack, F. W. Sabb, K. Foerde et al., “The neural cor-relates of motor skill automaticity,” Journal of Neuroscience,vol. 25, no. 22, pp. 5356–5364, 2005.

[89] T. Wu, K. Kansaku, and M. Hallett, “How self-initiatedmemorized movements become automatic: a functional MRIstudy,” Journal of Neurophysiology, vol. 91, no. 4, pp. 1690–1698, 2004.

[90] Y. Lajoie, N. Teasdale, C. Bard, and M. Fleury, “Attentionaldemands for static and dynamic equilibrium,” ExperimentalBrain Research, vol. 97, no. 1, pp. 139–144, 1993.

[91] Y. Lajoie, N. Teasdale, C. Bard, and M. Fleury, “Uprightstanding and gait: are there changes in attentional require-ments related to normal aging?” Experimental AgingResearch, vol. 22, no. 2, pp. 185–198, 1996.

[92] T. Wu and M. Hallett, “Neural correlates of dual taskperformance in patients with Parkinson’s disease,” Journal ofNeurology, Neurosurgery and Psychiatry, vol. 79, no. 7, pp.760–766, 2008.


[94] G. E. Alexander, M. R. DeLong, and P. L. Strick, “Parallelorganization of functionally segregated circuits linking basalganglia and cortex,” Annual Review of Neuroscience, vol. 9, pp.357–381, 1986.

[95] F. A. Middleton and P. L. Strick, “Anatomical evidence forcerebellar and basal ganglia involvement in higher cognitivefunction,” Science, vol. 266, no. 5184, pp. 458–461, 1994.

[96] F. A. Middleton and P. L. Strick, “Basal ganglia output andcognition: evidence from anatomical, behavioral, and clinicalstudies,” Brain and Cognition, vol. 42, no. 2, pp. 183–200,2000.

[97] D. J. Zgaljardic, J. C. Borod, N. S. Foldi, and P. Mattis, “Areview of the cognitive and behavioral sequelae of Parkinson’sdisease: relationship to frontostriatal circuitry,” Cognitive andBehavioral Neurology, vol. 16, no. 4, pp. 193–210, 2003.

[98] D. J. Zgaljardic, J. C. Borod, N. S. Foldi et al., “An examina-tion of executive dysfunction associated with frontostriatalcircuitry in Parkinson’s disease,” Journal of Clinical andExperimental Neuropsychology, vol. 28, no. 7, pp. 1127–1144,2006.

[99] P. Barone, “Neurotransmission in Parkinson’s disease:beyond dopamine,” European Journal of Neurology, vol. 17,no. 3, pp. 364–376, 2010.


[100] J. B. Leverenz, J. F. Quinn, C. Zabetian, J. Zhang, K. S.Montine, and T. J. Montine, “Cognitive impairment anddementia in patients with Parkinson disease,” Current Topicsin Medicinal Chemistry, vol. 9, no. 10, pp. 903–912, 2009.

[101] D. Devos, L. Defebvre, and R. Bordet, “Dopaminergicand non-dopaminergic pharmacological hypotheses for gaitdisorders in Parkinson’s disease,” Fundamental and ClinicalPharmacology, vol. 24, no. 4, pp. 407–421, 2010.

[102] C. Karachi, D. Grabli, F. A. Bernard et al., “Cholinergicmesencephalic neurons are involved in gait and postural dis-orders in Parkinson disease,” Journal of Clinical Investigation,vol. 120, no. 8, pp. 2745–2754, 2010.

[103] O. Blin, A. M. Ferrandez, J. Pailhous, and G. Serratrice,“Dopa-sensitive and Dopa resistant gait parameters inParkinson’s disease,” Journal of the Neurological Sciences, vol.103, no. 1, pp. 51–54, 1991.

[104] J. D. Schaafsma, N. Giladi, Y. Balash, A. L. Bartels, T.Gurevich, and J. M. Hausdorff, “Gait dynamics in Parkinson’sdisease: relationship to Parkinsonian features, falls andresponse to levodopa,” Journal of the Neurological Sciences,vol. 212, no. 1-2, pp. 47–53, 2003.

[105] I. C. Lopez, P. J. G. Ruiz, S. V. F. del Pozo, and V. S. Bernardos,“Motor complications in Parkinson’s disease: ten year follow-up study,” Movement Disorders, vol. 25, no. 16, pp. 2735–2739, 2010.

[106] P. J. Garcıa Ruiz, E. Meseguer, J. del Val, A. Vazquez, V.Sanchez Bernardos, and A. Vazquez, “Motor complicationsin Parkinson disease: a prospective follow-up study,” ClinicalNeuropharmacology, vol. 27, no. 2, pp. 49–52, 2004.

[107] S. J. G. Lewis, A. Slabosz, T. W. Robbins, R. A. Barker, andA. M. Owen, “Dopaminergic basis for deficits in workingmemory but not attentional set-shifting in Parkinson’sdisease,” Neuropsychologia, vol. 43, no. 6, pp. 823–832, 2005.

[108] R. Cools, R. A. Barker, B. J. Sahakian, and T. W. Robbins,“Enhanced or impaired cognitive function in Parkinson’sdisease as a function of dopaminergic medication and taskdemands,” Cerebral Cortex, vol. 11, no. 12, pp. 1136–1143,2001.

[109] M. Jahanshahi, L. Wilkinson, H. Gahir, A. Dharminda, andD. A. Lagnado, “Medication impairs probabilistic classifica-tion learning in Parkinson’s disease,” Neuropsychologia, vol.48, no. 4, pp. 1096–1103, 2010.

[110] R. J. S. George, J. G. Nutt, K. J. Burchiel, and F. B. Horak,“A meta-regression of the long-term effects of deep brainstimulation on balance and gait in PD,” Neurology, vol. 75,no. 14, pp. 1292–1299, 2010.

[111] A. J. Bastian, V. E. Kelly, F. J. Revilla, J. S. Perlmutter, andJ. W. Mink, “Different effects of unilateral versus bilateralsubthalamic nucleus stimulation on walking and reaching inParkinson’s disease,” Movement Disorders, vol. 18, no. 9, pp.1000–1007, 2003.

[112] M. Faist, J. Xie, D. Kurz et al., “Effect of bilateral subthalamicnucleus stimulation on gait in Parkinson’s disease,” Brain,vol. 124, part 8, pp. 1590–1600, 2001.

[113] V. E. Kelly, A. Samii, J. C. Slimp, R. Price, R. Goodkin, and A.Shumway-Cook, “Gait changes in response to subthalamicnucleus stimulation in people with Parkinson disease: a caseseries report,” Journal of Neurologic Physical Therapy, vol. 30,no. 4, pp. 184–194, 2006.

[114] D. Page and M. Jahanshahi, “Deep brain stimulation ofthe subthalamic nucleus improves set shifting but doesnot affect dual task performance in Parkinson’s disease,”IEEE Transactions on Neural Systems and RehabilitationEngineering, vol. 15, no. 2, pp. 198–206, 2007.

[115] J. L. Alberts, C. Voelcker-Rehage, K. Hallahan, M. Vitek, R.Bamzai, and J. L. Vitek, “Bilateral subthalamic stimulationimpairs cognitive—motor performance in Parkinson’s dis-ease patients,” Brain, vol. 131, part 12, pp. 3348–3360, 2008.



[118] E. J. Protas, K. Mitchell, A. Williams, H. Qureshy, K. Caroline,and E. C. Lai, “Gait and step training to reduce falls inParkinson’s disease,” NeuroRehabilitation, vol. 20, no. 3, pp.183–190, 2005.


[120] M. E. Morris, R. Iansek, and B. Kirkwood, “A randomizedcontrolled trial of movement strategies compared withexercise for people with Parkinson’s disease,” MovementDisorders, vol. 24, no. 1, pp. 64–71, 2009.


[122] D. K. White, R. C. Wagenaar, T. D. Ellis, and L. Tickle-Degnen, “Changes in walking activity and endurance follow-ing rehabilitation for people with Parkinson disease,” Archivesof Physical Medicine and Rehabilitation, vol. 90, no. 1, pp. 43–50, 2009.

[123] M. E. Morris, “Locomotor training in people with Parkinsondisease,” Physical Therapy, vol. 86, no. 10, pp. 1426–1435,2006.

[124] K. Baker, L. Rochester, and A. Nieuwboer, “The immediateeffect of attentional, auditory, and a combined cue strategyon gait during single and dual tasks in Parkinson’s disease,”Archives of Physical Medicine and Rehabilitation, vol. 88, no.12, pp. 1593–1600, 2007.

[125] C. G. Canning, “The effect of directing attention duringwalking under dual-task conditions in Parkinson’s disease,”Parkinsonism and Related Disorders, vol. 11, no. 2, pp. 95–99,2005.

[126] M. E. Morris, R. Iansek, T. A. Matyas, and J. J. Summers,“Stride length regulation in Parkinson’s disease: normaliza-tion strategies and underlying mechanisms,” Brain, vol. 119,part 2, pp. 551–568, 1996.

[127] L. Rochester, V. Hetherington, D. Jones et al., “The effectof external rhythmic cues (auditory and visual) on walkingduring a functional task in homes of people with Parkinson’sdisease,” Archives of Physical Medicine and Rehabilitation, vol.86, no. 5, pp. 999–1006, 2005.

[128] L. Rochester, K. Baker, V. Hetherington et al., “Evidencefor motor learning in Parkinson’s disease: acquisition,automaticity and retention of cued gait performance aftertraining with external rhythmical cues,” Brain Research, vol.1319, pp. 103–111, 2010.

[129] B. Hu, N. de Bruin, J. B. Doan et al., “Walking with music isa safe and viable tool for gait training in parkinson’s disease:


the effect of a 13-week feasibility study on single and dual taskwalking,” Parkinson’s Disease, vol. 2010, Article ID 483530, 9pages, 2010.

[130] L. Rochester, D. Rafferty, C. Dotchin, O. Msuya, V. Minde,and R. W. Walker, “The effect of cueing therapy on singleand dual-task gait in a drug naıve population of peoplewith parkinson’s disease in Northern Tanzania,” MovementDisorders, vol. 25, no. 7, pp. 906–911, 2010.

[131] L. Rochester, D. J. Burn, G. Woods, J. Godwin, and A.Nieuwboer, “Does auditory rhythmical cueing improve Gaitin people with Parkinson’s disease and cognitive impairment?A feasibility study,” Movement Disorders, vol. 24, no. 6, pp.839–845, 2009.

[132] A. L. Behrman, P. Teitelbaum, and J. H. Cauraugh, “Verbalinstructional sets to normalise the temporal and spatialgait variables in Parkinson’s disease,” Journal of NeurologyNeurosurgery and Psychiatry, vol. 65, no. 4, pp. 580–582,1998.

[133] C. A. Lohnes and G. M. Earhart, “The impact of attentional,auditory, and combined cues on walking during single andcognitive dual tasks in Parkinson disease,” Gait and Posture,vol. 33, no. 3, pp. 478–483, 2011.

[134] P. Fok, M. Farrell, and J. McMeeken, “Prioritizing gait indual-task conditions in people with Parkinson’s,” HumanMovement Science, vol. 29, no. 5, pp. 831–842, 2010.

[135] P. Fok, M. Farrell, and J. McMeeken, “The effect of dividingattention between walkingand auxiliary tasks in people withParkinson’s disease,” Human Movement Science. In press.

[136] C. G. Canning, L. Ada, and E. Woodhouse, “Multiple-taskwalking training in people with mild to moderate Parkinson’sdisease: a pilot study,” Clinical Rehabilitation, vol. 22, no. 3,pp. 226–233, 2008.

[137] S. G. Brauer and M. E. Morris, “Can people with Parkinson’sdisease improve dual tasking when walking?” Gait andPosture, vol. 31, no. 2, pp. 229–233, 2010.

[138] O. Beauchet, V. Dubost, G. Allali, R. Gonthier, F. R. Her-mann, and R. W. Kressig, “’Faster counting while walking’as a predictor of falls in older adults,” Age and Ageing, vol. 36,no. 4, pp. 418–423, 2007.

[139] O. Beauchet, G. Allali, C. Annweiler et al., “Does change ingait while counting backward predict the occurrence of a firstfall in older adults?” Gerontology, vol. 54, no. 4, pp. 217–223,2008.

[140] A. Bootsma-van der Wiel, J. Gussekloo, A. J. M. de Craen,E. van Exel, B. R. Bloem, and R. G. J. Westendorp, “Walkingand talking as predictors of falls in the general population:the Leiden 85-plus study,” Journal of the American GeriatricsSociety, vol. 51, no. 10, pp. 1466–1471, 2003.

[141] L. Lundin-Olsson, L. Nyberg, and Y. Gustafson, “’Stopswalking when talking’ as a predictor of falls in elderly people,”Lancet, vol. 349, no. 9052, p. 617, 1997.

[142] K. L. McCulloch, V. Mercer, C. Giuliani, and S. Marshall,“Development of a clinical measure of dual-task performancein walking: reliability and preliminary validity of the walkingand remembering test,” Journal of Geriatric Physical Therapy,vol. 32, no. 1, pp. 2–9, 2009.

[143] A. Shumway-Cook, S. Brauer, and M. Woollacott, “Predict-ing the probability for falls in community-dwelling olderadults using the Timed Up and Go Test,” Physical Therapy,vol. 80, no. 9, pp. 896–903, 2000.

[144] J. Verghese, H. Buschke, L. Viola et al., “Validity of dividedattention tasks in predicting falls in older individuals: apreliminary study,” Journal of the American Geriatrics Society,vol. 50, no. 9, pp. 1572–1576, 2002.

[145] B. R. Bloem, Y. A. M. Grimbergen, M. Cramer, and V. V.Valkenburg, “’Stops walking when talking’ does not predictfalls in Parkinson’s disease,” Annals of Neurology, vol. 48, no.2, p. 268, 2000.

[146] A. Mirelman, I. Maidan, T. Herman, J. E. Deutsch, N. Giladi,and J. M. Hausdorff, “Virtual reality for gait training: canit induce motor learning to enhance complex walking andreduce fall risk in patients with Parkinson’s disease?” Journalsof Gerontology, Series A, vol. 66, no. 2, pp. 234–240, 2011.


Research Article

Reliability in One-Repetition Maximum Performance inPeople with Parkinson’s Disease

Thomas A. Buckley1 and Christopher J. Hass2

1 Department of Health and Kinesiology, Georgia Southern University, Statesboro, GA 30460, USA2 Department of Applied Physiology and Kinesiology and Movement Disorders Center, University of Florida,Gainesville, FL 32611, USA

Correspondence should be addressed to Thomas A. Buckley, [email protected]

Received 30 May 2011; Revised 5 August 2011; Accepted 1 September 2011

Academic Editor: Lee Dibble

Copyright © 2012 T. A. Buckley and C. J. Hass. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

Strength training is an effective modality to improve muscular strength and functional performance in people with Parkinson’sdisease (PWP). One-repetition maximum (1-RM) is the gold standard assessment of strength; however, PWP suffer from day-to-day variations in symptom severity and performance characteristics, potentially adversely affecting the reliability of 1-RMperformance. Herein, we assessed the reliability of 1-RM in PWP. Forty-six participants completed two sessions of 1-RM testing ofknee extension, knee flexion, chest press, and biceps curl at least 72 hours apart. Significantly differences between testing sessionswere identified for knee extension (P < 0.001), knee flexion (P = 0.042), and biceps curl (P = 0.001); however, high reliability (ICC>0.90) was also identified between sessions. Interestingly, almost third of subjects failed to perform better on the second testingsession. These findings suggest that 1-RM testing can be safely performed in PWP and that disease-related daily variability mayinfluence 1-RM performance.

1. Introduction

Parkinson’s disease (PD), a progressive neurological diseasewhich is believed to affect over 1.5 million Americans, resultsfrom the degeneration of the dopaminergic neurons in themidbrain and the resulting reduced dopamine availabilityto the basal ganglia [1, 2]. The cardinal features of PDinclude rigidity, tremor, bradykinesia, and impaired posturalcontrol, and these symptoms are often unpredictable andtheir severity can fluctuate daily, often termed “day-to-dayvariability” [3–5]. Further, muscular weakness, identifiedby Dr. Parkinson as an early symptom of the disease, isalso frequently reported by people with Parkinson’s (PWP)[6, 7]. However, inconsistent findings in the literature haveobscured the elucidation of the underlying mechanismof the apparent weakness, thus, raising the debate ifmuscular weakness is intrinsic to the disease or a secondaryconsequence [8, 9]. Muscular weakness, when present inPWP, presents bilaterally and tends to increase as the velocity

of movement increases [9]. While the specific contributoryneurophysiological mechanisms remain uncertain, bradyki-nesia, the inability to energize the appropriate muscles togenerate forces at a sufficient rate, is thought to be a majorcontributing factor [8, 10]. Bradykinesia likely results frombasal ganglia pathophysiology leading to impairments inboth motor programming and execution [11]. Muscularweakness and bradykinesia impair power production,particularly at lighter loads [8]. These reductions in muscularstrength and power have been associated with both reducedfunctional ambulation and impaired dynamic posturalstability in PWP [12–14]. As a result many patients with PDreceive physical therapy services to counteract these deficits.

Recent reviews have suggested that strength training maybe an effective modality to improve strength and functionalperformance for PWP [15, 16]. Strength training hasfrequently been combined with other rehabilitative proto-cols including cueing strategies, aerobic or cardiovasculartraining, balance training, stretching exercises, and creatine


supplementation in the development of global rehabilitationprograms [17–25]. These programs have led to increasedmuscular strength [17–20], reduced bradykinesia [21], andimproved cognitive functioning [22, 23]. Further, theseimprovements have transferred to overall increased quality oflife [21, 25] and improved functional performance includinggait [26], sit to stand [27, 28], sit to walk [29], and overallfunctional mobility [18]. It is not surprising, therefore, thatstrength training programs have become more integratedinto successful Parkinson rehabilitation programs.

An important first step in initiating a rehabilitationprogram is the assessment of baseline function by whichtherapy-based improvements can be judged. When resistancetraining is a component of the therapeutic protocol, assess-ment of baseline strength is paramount. Though multipleoptions exist, including more subjective manual muscletesting, the accepted gold standard of maximal muscle testingis the use of the one-repetition maximum (1-RM) test [30].The 1-RM is defined as the maximal weight that can belifted once with correct lifting technique and is generallyconsidered to have good to excellent (ICC > 0.95) reliabilityin healthy adults [31, 32]. However, therapists and rehabil-itation specialists need to be aware of the determinants of1-RM testing which include both previous weight trainingexperience and familiarization with the test [33–35]. Furtherchallenging the assessment of muscular performance aredisease-specific complications including the prevalent motorfluctuations, random changes in symptoms severity, andnoted “on/off” daily variability [36–38].

Previous rehabilitation studies in PWP have utilizedeither one or two sessions of various strength testing proto-cols to identify the individual’s current strength; however, thereliability of these protocols, specifically maximal strengthassessment, has not been assessed in this population [17,20, 24, 26]. Therefore, the purpose of this study wasto investigate the reliability of 1-RM testing in mild-to-moderate PWP across two testing sessions. We hypothesizedthat 1-RM testing would be generally reliable; however, thedisease related day-to-day variability associated with PDwould result in individuals differences during the testing.

2. Methods

2.1. Subjects. A total of 46 participants diagnosed with idio-pathic PD by a movement disorder neurologist participatedin this study (Table 1). Inclusion criteria included a modifiedHoehn and Yahr stage 1–3, the ability to ambulate withoutassistance, and stable response to anti-Parkinson medica-tions. Exclusion criteria included cardiovascular, musculos-ketal, vestibular disorders, or other neurological conditionsbeyond PD or recent enrollment in an exercise trainingprogram. All participants were tested while clinically “on”approximately 1–1.5 hours following the first medicationdose of the day and self-reported that their medicines wereworking maximally at the time of testing. No participantsdemonstrated any dyskinesia or freezing during the testingsessions. All participants provided written informed consentprior to participating in the study as approved by theUniversity’s Institutional Review Board.

Table 1: Participant demographics and anthropometric data.Anthropometric data is presented as mean ± standard deviation.Hoehn and Yahr classification is presented as the actual number ofsubjects and the percentage of the total (percentage does not add to100% due to rounding).

Participant characteristics

Age (years) 62.6 ± 4.8

Height (m) 1.72 ± 0.11

Weight (kg) 86.8 ± 13.8

Disease duration (years) 10.9 ± 9.9

Hoehn & Yahr score 2.3 ± 0.6

Hoehn & Yahr 1 2 (4.3%)

Hoehn & Yahr 1.5 7 (15.2%)

Hoehn & Yahr 2 14 (30.4%)

Hoehn & Yahr 2.5 11 (23.9%)

Hoehn & Yahr 3 12 (26.1%)

Unified Parkinson Disease Rating Scale (UPDRS)∗

Total score 38.0 ± 6.1

Motor score 23.8 ± 4.6

ADL score 12.2 ± 2.2∗

UPDRS data was only available on 25 of the 46 subjects.

2.2. Experimental Procedures. Prior to performing the 1-RM testing sessions, all participants underwent two famil-iarization sessions, between 48–72 hours apart, to orientatethemselves with the exercise equipment. During these ses-sions the appropriate positioning and lifting techniques wereinstructed and each subject performed two sets of eachexercise at a low-to-moderate resistance level. The followingweek, the 1-RM tests were performed using cable-loadedresistance machines for knee extension (KE), knee flexion(KF) (New York Barbell, Elmira, NY.), chest press (CP), andbiceps curl (BC) (Nautilus Corp, Vancouver, WA.). Both the1-RM testing protocol and the participants body alignmentfor each tested closely adhered to the recommendations ofthe National Strength and Conditioning Association [30].For each exercise, subjects warmed up with a low resistanceand performed 10 repetitions. Thereafter, resistance wasincreased in incremental loads until failure occurred despiteverbal encouragement to continue [17]. In order to beclassified as a successful attempt, the subject had to move theweight through the complete range of motion in a controlledmanner without compensatory movements (e.g., shiftingbody position). The 1-RM was determined within 5 attemptsfor all subjects.

In order to reduce the potential confounding effects offatigue, no individual performed more than two 1-RM testsin a given day and at least 72 hours rest was provided betweentests. Specifically, on a given test day the subject wouldperform one upper body and one lower body assessment.All 46 subjects performed the KE 1-RM tests, followed by 25subjects performing the BC, 24 subjects performing the CP,and 21 subjects performing the KF.

2.3. Statistical Analysis. The same investigator tested theparticipants on both days. A paired sample T-test was


Table 2: One-repetition maximum test results. The session 1-RM values are presented as mean ± standard deviation.

ExerciseFirst session

(kg)Second session

(kg)Mean session difference

(kg) (95% CI)T-test results ICC (95% CI) SEM

Knee extension 63.7 ± 28.1 67.7 ± 29.7 4.0 (1.9–6.2) P < 0.001 .96 (.93–.97) 5.7 kg

Knee flexion 27.0 ± 12.7 29.4 ± 13.0 2.4 (0.2–4.7) P = 0.042 .91 (.79–.96) 3.8 kg

Biceps curl 43.9 ± 15.6 46.6 ± 17.6 2.7 (1.2–4.1) P = 0.001 .97 (.92–.98) 2.8 kg

Chest press 57.8 ± 20.6 60.1 ± 20.8 2.3 (−0.2–4.7) P = 0.066 .95 (.90–.98) 4.3 kg

ICC: Intraclass Correlation Coefficient. SEM: Standard Error of Measurement which was calculated as: SEM: SDbaseline ∗√(1− rtest–retest).

performed to compare differences between 1-RM duringsession 1 and session 2 for each of the four exercises. Themean difference and 95% confidence intervals between thetwo tests were calculated as session 2 minus session 1,such that a positive number indicates an increase in 1-RMduring session 2. A frequency distribution was performed foreach exercise to identify which test session most commonlyrepresented the higher value. The intraclass correlationcoefficient (ICC) was calculated for each exercise with a two-way random effects analysis of variance. Finally, the standarderror of the measurement (SEM) was calculated as SEM =SDbaseline ∗√(1− rtest−retest) [39].

3. Results

All subjects completed all 1-RM tests without incident. Thepaired analysis revealed statistically significant differences in1-RM performance between the two testing sessions for kneeextension, knee flexion, and biceps curl, but not for chestpress (Table 2). The intraclass correlation coefficient rangedfrom 0.91 to 0.97 (Table 2).

Across the four exercises, a total of 116 tests wereperformed; of these, 11.2% (13 of 116) had identical scoresbetween the two testing sessions. Further, 19.8% (23 of 116)of the evaluations had higher 1-RM values, a mean of 4.6 kgacross all 4 exercises, on the first test. Finally, the rangeof differences between the two testing sessions was 82% ofthe combined means (54 kg) with one participant increasingtheir 1-RM by 41% (27 kg) and another subject exhibiting a41% (27 kg) reduction in 1-RM, both occurred during kneeextension exercises, and over half of all participants (51%)had changes of at least 5 kg between test sessions.

4. Discussion

Effective and reliable assessment of force production is anintegral component in the development of an appropriatephysical therapy program. Further, in longitudinal studies itis essential to establish an accurate and reliable baseline per-formance of strength to compare improvements over time.The purpose of this study was to investigate reliability in 1-RM performance amongst PWP. A primary finding of thisstudy was a significant difference in 1-RM strength betweenthe two sessions for knee extension, knee flexion, andbiceps curl exercises in individuals with mild-to-moderatePD despite the subjects performing two orientation sessions

in the previous week. However, the tests demonstrated highreliability and the between sessions differences did not exceedthe standard error of measurement when collapsed acrossparticipants. Interestingly, nearly third of subjects did notincrease their 1-RM on the second testing session as wouldbe expected in this inexperienced population. In some cases,the improvements we observed (up to 41% improvement)rival or exceed those reported in many longitudinal trainingstudies [17, 18, 20, 21]. This finding suggests that day-to-dayperformance variability may play a substantial role in 1-RMstrength testing for individuals with mild-to-moderate PD.

Accurate and reliable baseline testing needs to be con-ducted to correctly prescribe the treatment protocol andelucidate improvements following exercise programs. Theresults of this study suggest that more than one baseline 1-RM test needs to be performed, although therapists shouldnot assume improved performance with second-day testing.Indeed, over 30% of subjects failed to improve in 1-RMperformance on the second testing session and a between-test range of 54 kg was identified during the leg extensionexercise. This finding raises two unique concerns to thedevelopment and reporting on the effects of strengtheningprograms for Parkinson’s rehabilitation. First, if the initial1-RM value is low, the exercise prescription based on thisvalue may not be sufficiently challenging to the individual,thus, potentially limiting the effectiveness of the therapy.Secondly, variable performance raises the risk that the truebenefit of the intervention may be masked by a single daypoor performance in a population known to experience day-to-day performance variability [5, 40, 41]. The results ofthis study are similar to recent finding of aerobic capacityin PWP [42]. Katzel and colleagues demonstrated generallyhigh test-retest reliability, however a significant betweentest session, 0.56 mL/mg/min, difference was noted in VO2

peak measurements [42]. Further, almost half of the PWP,failed to improve on the second administration of themaximal test (95% CI of −3.5–4.6 mL/mg/min) [42]. Takentogether, these findings provide important considerations inthe development of rehabilitation programs for individualswith mild-to-moderate PD.

The phenomenon of day-to-day variability in PWP hasbeen well established in the literature [5, 40, 41, 43]. Thesymptoms of Parkinson’s, both physical and psychological,are often unpredictable and fluctuate from day to day result-ing in substantial alterations in activities of daily living andsocial activities [40, 44]. This is a separate phenomenon frommotor fluctuations, abrupt and unpredictable responses to


levodopa administration [45]. Further, both hourly and dailyvariations, potentially due to motor fluctuations or day-to-day variations, in gait rhythm (e.g., velocity, step length, andcadence), have been identified in PWP [46]. The participantsin this study were all tested at a consistent time followingmedication dosage, at their self-described best time of day,and while clinically “on”; so only subtle motor fluctuationscould have been a contributing factor to their performance.

The use of 1-RM testing has been examined in a widerange of healthy, aging, and diseased populations [32, 35,47–54]. In healthy young adults (age 18–30) with strengthtraining experience, the reliability of the 1-RM test isgenerally considered to be very high (ICC > 0.95) [47,55]. In healthy older adults, individuals with cardiovasculardisease, peripheral obstructive arterial disease, and chronicobstructive pulmonary disease, 1-RM testing is a safe andpractical assessment and our results suggest 1-RM testingis also safe amongst the PWP population with comparablereliability [35, 52–54]. Interestingly, Schilling et al. [20]recently found no differences in maximal relative strengthtesting, reported as maximum strength divided by bodyweight; however, these tests were separated by 8 weeks, asopposed to 72 hours, and the time between tests may haveinfluenced the relation to our results. The results of thecurrent study suggest that PWP can safely and effectivelyperform 1-RM testing and, while important differences existbetween trials, the overall results are generally reliable.

Generally speaking, the reliability of 1-RM measures mayvary depending on the individuals experience with weighttraining and their familiarity with the specific exercise beingtested [32, 33, 47–49, 55]. Although the number of acceptablefamiliarization sessions has ranged from one to nine, inhealthy inexperienced middle-aged to older populations, oneto three familiarization sessions are generally considered tobe appropriate before assessing maximal strength [32, 34,35]. Following familiarization with the equipment, moststudies on healthy older adults suggest that two to three1-RM sessions are required as strength values will increaseon subsequent trials [33–35]. While the specific mechanismunderlying these improvements in 1-RM performance, whenpresent, is not fully understood, it is generally attributedto improved neural efficiency and activation patterns aswell as a learning effect represented by improved postureand exercise execution [33, 56]. Appropriate orientation andfamiliarization to the testing paradigm is likely of particularimportance for PWP who are known to reduce overallactivity due to social stigmas, loss of confidence in theircoordination, and fear of falling [26, 57].

The findings of this study are delimited to this specificprotocol, and future studies should address this potentiallimitation by increasing the number of both familiarizationand 1-RM testing sessions to help elucidate the learningeffects and the influence of day-to-day variability. Further,additional demographic considerations (e.g., UPDRS scores)and traditional performance variables (e.g., timed get-upand go test) should be explored to identify potential rela-tionships. However, exploratory analysis of our data foundno relationship between disease severity as measured byHoehn and Yahr staging, body weight or initial strength, and

the change in performance between testing sessions. Whileday-to-day variability in PWP is unpredictable, exerciseintervention studies should consider a Parkinson’s specificgraded symptom checklist on the days of the pre- andposttesting to attempt to control for the variability. Finally,future studies should expand these findings by identifyingpotential relationships between alterations in strength andperformance of activities of daily living.

The 1-RM test is generally considered to be the goldstandard for assessing maximal muscular strength in an indi-vidual and the results of this study suggest that, when usingcable-loaded resistance machines, PWP can successfullyand safely perform these tests [30]. Thus, physical therapyinterventions can effectively be established and monitoredwith 1-RM testing in the PD population. Whereas healthyolder adults typically demonstrate subtle improvements in1-RM performance with repeat administration over severaldays, the results of this study suggest that individuals withmild-to-moderate PD demonstrate inconsistencies in 1-RMtest performance.

References

[1] T. Hampton, “Parkinson disease registry launched,” Journal ofthe American Medical Association, vol. 293, no. 2, p. 149, 2005.

[2] A. Galvan and T. Wichmann, “Pathophysiology of parkinson-ism,” Clinical Neurophysiology, vol. 119, no. 7, pp. 1459–1474,2008.

[3] D. B. Calne, B. J. Snow, and C. Lee, “Criteria for diagnosingParkinson’s disease,” Annals of Neurology, vol. 32, pp. S125–S127, 1992.

[4] C. C. Goodman and W. G. Boissonnault, Pathology: Implica-tions for the Physical Therapist, WB Saunders, Philadelphia, Pa,USA, 3rd edition, 1998.

[5] G. A. Dakof and G. A. Mendelsohn, “Parkinson’s disease.The psychological aspects of a chronic illness,” PsychologicalBulletin, vol. 99, no. 3, pp. 375–387, 1986.

[6] J. Parkinson, An Essay on the Shaking Palsy, Whittingham andRoland, London, UK, 1817.

[7] M. Stacy, A. Bowron, M. Guttman et al., “Identificationof motor and nonmotor wearing-off in Parkinson’s disease:comparison of a patient questionnaire versus a clinicianassessment,” Movement Disorders, vol. 20, no. 6, pp. 726–733,2005.

[8] N. E. Allen, C. G. Canning, C. Sherrington, and V. S. C. Fung,“Bradykinesia, muscle Weakness and reduced muscle powerin Parkinson’s disease,” Movement Disorders, vol. 24, no. 9, pp.1344–1351, 2009.

[9] R. Cano-De-La-Cuerda, M. Perez-De-Heredia, J. C.Miangolarra-Page, E. Munoz-Hellın, and C. Fernandez-De-Las-Penas, “Is there muscular weakness in parkinson’sdisease?” American Journal of Physical Medicine and Rehabil-itation, vol. 89, no. 1, pp. 70–76, 2010.



[12] J. R. Nocera, T. Buckley, D. Waddell, M. S. Okun, and C. J.Hass, “Knee extensor strength, dynamic stability, and func-tional ambulation: are they related in Parkinson’s disease?”


Archives of Physical Medicine and Rehabilitation, vol. 91, no.4, pp. 589–595, 2010.

[13] L. M. Inkster, J. J. Eng, D. L. MacIntyre, and A. Jon Stoessl,“Leg muscle strength is reduced in Parkinson’s disease andrelates to the ability to rise from a chair,” Movement Disorders,vol. 18, no. 2, pp. 157–162, 2003.

[14] B. K. Schilling, R. E. Karlage, M. S. LeDoux, R. F. Pfeiffer, L.W. Weiss, and M. J. Falvo, “Impaired leg extensor strengthin individuals with Parkinson disease and relatedness tofunctional mobility,” Parkinsonism and Related Disorders, vol.15, no. 10, pp. 776–780, 2009.



[17] C. J. Hass, M. A. Collins, and J. L. Juncos, “Resistance trainingwith creatine monohydrate improves upper-body strengthin patients with Parkinson disease: a randomized trial,”Neurorehabilitation and Neural Repair, vol. 21, no. 2, pp. 107–115, 2007.

[18] L. E. Dibble, T. F. Hale, R. L. Marcus, J. Droge, J. P. Gerber,and P. C. LaStayo, “High-intensity resistance training amplifiesmuscle hypertrophy and functional gains in persons withparkinson’s disease,” Movement Disorders, vol. 21, no. 9, pp.1444–1452, 2006.

[19] T. Toole, M. A. Hirsch, A. Forkink, D. A. Lehman, and C.G. Maitland, “The effects of a balance and strength trainingprogram on equilibrium in Parkinsonism: a preliminarystudy,” NeuroRehabilitation, vol. 14, no. 3, pp. 165–174, 2000.

[20] B. K. Schilling, R. F. Pfeiffer, M. S. Ledoux, R. E. Karlage, R. J.Bloomer, and M. J. Falvo, “Effects of moderate-volume, high-load lower-body resistance training on strength and functionin persons with parkinson’s disease: a pilot study,” Parkinson’sDisease, Article ID 824734, 2010.


[22] K. Tanaka, A. C. D. Quadros, R. F. Santos, F. Stella, L. T.B. Gobbi, and S. Gobbi, “Benefits of physical exercise onexecutive functions in older people with Parkinson’s disease,”Brain and Cognition, vol. 69, no. 2, pp. 435–441, 2009.

[23] K. E. Cruise, R. S. Bucks, A. M. Loftus, R. U. Newton,R. Pegoraro, and M. G. Thomas, “Exercise and Parkinson’s:benefits for cognition and quality of life,” Acta NeurologicaScandinavica, vol. 123, no. 1, pp. 13–19, 2011.


[25] F. Rodrigues de Paula, L. F. Teixeira-Salmela, C. D. Coelho deMorais Faria, P. Rocha de Brito, and F. Cardoso, “Impact ofan exercise program on physical, emotional, and social aspectsof quality of life of individuals with Parkinson’s disease,”Movement Disorders, vol. 21, no. 8, pp. 1073–1077, 2006.

[26] T. A. Scandalis, A. Bosak, J. C. Berliner, L. L. Helman,and M. R. Wells, “Resistance training and gait function in

patients with Parkinson’s disease,” American Journal of PhysicalMedicine and Rehabilitation, vol. 80, no. 1, pp. 38–43, 2001.

[27] P. K. Schot, K. M. Knutzen, S. M. Poole, and L. A. Mrotek,“Sit-to-stand performance of older adults following strengthtraining,” Research Quarterly for Exercise and Sport, vol. 74, no.1, pp. 1–8, 2003.


[29] T. A. Buckley, C. Pitsikoulis, and C. J. Hass, “Dynamic posturalstability during sit-to-walk transitions in Parkinson diseasepatients,” Movement Disorders, vol. 23, no. 9, pp. 1274–1280,2008.

[30] T. R. Baechle and R. W. Earle, Essentials of Strength Trainingand Conditioning, Human Kinetics Books, Champaign, Ill,USA, 3rd edition, 2008.

[31] ACSM, ACSM’s Resource Manual for Guidelines for ExerciseTesting and Prescription, Lippincott Williams & Wilkins,Philadelphia, Pa, USA, 6th edition, 2009.

[32] I. Levinger, C. Goodman, D. L. Hare, G. Jerums, D. Toia, andS. Selig, “The reliability of the 1RM strength test for untrainedmiddle-aged individuals,” Journal of Science and Medicine inSport, vol. 12, no. 2, pp. 310–316, 2009.

[33] R. M. Ritti-Dias, A. Avelar, E. P. Salvador, and E. S. Cyrino,“Influence of previous experience on resistance training onreliability of one-repetition maximum test,” Journal of Strengthand Conditioning Research, vol. 25, no. 5, pp. 1418–1422, 2011.

[34] L. L. Ploutz-Snyder and E. L. Giamis, “Orientation andFamiliarization to 1RM Strength Testing in Old and YoungWomen,” Journal of Strength and Conditioning Research, vol.15, no. 4, pp. 519–523, 2001.

[35] W. T. Phillips, A. M. Batterham, J. E. Valenzuela, and L. N.Burkett, “Reliability of Maximal Strength Testing in OlderAdults,” Archives of Physical Medicine and Rehabilitation, vol.85, no. 2, pp. 329–334, 2004.

[36] M. MacKay-Lyons, “Variability in spatiotemporal gait charac-teristics over the course of the L-dopa cycle in people withadvanced parkinson disease,” Physical Therapy, vol. 78, no. 10,pp. 1083–1094, 1998.

[37] J. Reimer, M. Grabowski, O. Lindvall, and P. Hagell, “Use andinterpretation of on/off diaries in Parkinson’s disease,” Journalof Neurology, Neurosurgery and Psychiatry, vol. 75, no. 3, pp.396–400, 2004.

[38] B. Lindsey, “Hourly monitoring system for patients withParkinson’s disease,” Neurology Report, vol. 19, no. 1, pp. 30–33, 1995.

[39] J. P. Weir, “Quantifying test-retest reliability using the intra-class correlation coefficient and the SEM,” Journal of Strengthand Conditioning Research, vol. 19, no. 1, pp. 231–240, 2005.

[40] K. Shifren, P. Wood, K. Hooker, and J. R. Nesselroade, “Struc-ture and variation of mood in individuals with Parkinson’sdisease: a dynamic factor analysis,” Psychology and Aging, vol.12, no. 2, pp. 328–339, 1997.

[41] J. J. van Hilten, H. A. M. Middelkoop, G. A. Kerkhof, and R. A.C. Roos, “A new approach in the assessment of motor activityin Parkinson’s disease,” Journal of Neurology Neurosurgery andPsychiatry, vol. 54, no. 11, pp. 976–979, 1991.

[42] L. I. Katzel, J. D. Sorkin, R. F. Macko, B. Smith, F. M.Ivey, and L. M. Shulman, “Repeatability of aerobiccapacitymeasurements in Parkinson disease,” Medicine & Science inSports & Exercise. In press.

[43] P. D. Society, Fatigue and Parkinson’s, 2008.


[44] M. Smeja, F. Foerster, G. Fuchs, D. Emmans, A. Hornig, and J.Fahrenberg, “24-h assessment of tremor activity and posturein Parkinson’s disease by multi-channel accelerometry,” Jour-nal of Psychophysiology, vol. 13, no. 4, pp. 245–256, 1999.

[45] J. G. Nutt, “On-off phenomenon: relatin to levodopa pharma-cokinetics and pharmacodynamics,” Annals of Neurology, vol.22, no. 4, pp. 535–540, 1987.

[46] H. Mitoma, M. Yoneyama, and S. Orimo, “24-hour recordingof parkinsonian gait using a portable gait rhythmogram,”Internal Medicine, vol. 49, no. 22, pp. 2401–2408, 2010.

[47] W. J. Kraemer, N. Ratamess, A. C. Fry et al., “Influence ofresistance training volume and periodization on physiologicaland performance adaptations in collegiate women tennisplayers,” American Journal of Sports Medicine, vol. 28, no. 5,pp. 626–633, 2000.

[48] W. Kroll, “Reliability of a selected measure of human strength,”Research Quarterly, vol. 33, no. 3, pp. 410–417, 1962.

[49] W. Kroll, “Reliability variations of strength in test-retestsituations,” Research Quarterly, vol. 34, no. 1, pp. 50–55, 1963.

[50] E. T. Schroeder, Y. Wang, C. Castaneda-Sceppa et al., “Relia-bility of maximal voluntary muscle strength and power testingin older men,” Journals of Gerontology. Series A, vol. 62, no. 5,pp. 543–549, 2007.

[51] M. J. Benton, P. D. Swan, and M. D. Peterson, “Evaluation ofmultiple one repetition maximum strength trials in untrainedwomen,” Journal of Strength and Conditioning Research, vol.23, no. 5, pp. 1503–1507, 2009.

[52] R. M. R. Dias, G. G. Cucato, L. C. Camara, and N. Wolosker,“Reproducibility of the 1-RM test in individuals with periph-eral obstructive arterial disease,” Revista Brasileira de Medicinado Esporte, vol. 16, no. 3, pp. 201–204, 2010.

[53] K. L. Barnard, K. J. Adams, A. M. Swank, E. Mann, andD. M. Denny, “Injuries and muscle soreness during the onerepetition maximum assessment in a cardiac rehabilitationpopulation,” Journal of Cardiopulmonary Rehabilitation, vol.19, no. 1, pp. 52–58, 1999.

[54] V. S. Dourado, S. E. Tanni, L. C. O. Antunes et al., “Effect ofthree exercise programs on patients with chronic obstructivepulmonary disease,” Brazilian Journal of Medical and BiologicalResearch, vol. 42, no. 3, pp. 263–271, 2009.

[55] M. R. Rhea, S. D. Ball, W. T. Phillips, and L. N. Burkett,“A comparison of linear and daily undulating periodizedprograms with equated volume and intensity for strength,”Journal of Strength and Conditioning Research, vol. 16, no. 2,pp. 250–255, 2002.

[56] R. M. Enoka, “Neural adaptations with chronic physicalactivity,” Journal of Biomechanics, vol. 30, no. 5, pp. 447–455,1997.



Research Article

Comparing the Mini-BESTest with the Berg Balance Scale toEvaluate Balance Disorders in Parkinson’s Disease

Laurie A. King,1 Kelsey C. Priest,1 Arash Salarian,1 Don Pierce,2 and Fay B. Horak1

1 Department of Neurology, Oregon Health & Science University, Portland, OR 97239, USA2 Division of Biostatistics, Department of Public Health and Preventive Medicine,Oregon Health & Science University, Portland, OR 97239, USA

Correspondence should be addressed to Laurie A. King, [email protected]

Received 9 June 2011; Revised 23 July 2011; Accepted 18 August 2011


Copyright © 2012 Laurie A. King et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Objective. The purpose of this study was to explore the usefulness of the Mini-BESTest compared to the Berg Balance Scale inevaluating balance in people with PD of varying severity. We evaluated (1) the distribution of patients scores to look for ceilingeffects, (2) concurrent validity with severity of disease, and (3) the sensitivity/specificity of separating people with or withoutpostural response deficits. Subjects. Ninety-seven people with PD were tested for balance deficits using the Berg, Mini-BESTest,Unified Parkinson’s Disease Rating Scale (UPDRS) III and the Hoehn & Yahr (H&Y) disease severity classification. Setting. Clinicalresearch facility at Oregon Health & Science University. Results. The Mini-BESTest is highly correlated with the Berg (r = 0.79,P < 0.001), but avoids the ceiling compression effect of the Berg for mild PD (skewness −2.30 Berg, −0.93 Mini-BESTest).Consequently, the Mini-BESTest is more effective than the Berg for predicting UPDRS Motor score (P < 0.001 Mini-BESTestversus P = 0.86 Berg), and for discriminating between those with and without postural response deficits as measured by the H&Y(ROC differential P = 0.06). Conclusion. The Mini-BESTest is a promising tool for discerning balance deficits in patients with PD,most importantly those with more subtle deficits.

1. Introduction

Postural instability and balance deficits are one of the mostdebilitating impairments associated with chronic neuro-logical disease, such as Parkinson’s disease (PD) [1]. Themost commonly used clinical test of balance severity inpeople with PD is the Berg Balance Scale (Berg) [2]. TheBerg, originally designed for use in the frail elderly, isa 14-item test that focuses on a variety of self-initiatedtasks related to everyday function such as sit-to-stand andfunctional reach forward. The Berg has excellent reliabilityand is somewhat correlated with severity of PD, as measuredwith the Unified Parkinson Rating Scale (UPDRS) [3, 4].However, the Berg has limitations such as documentedceiling effects [5–7] and problems with underutilization andredundancy of categories due to the rating scale [8, 9]. Theseparticular limitations are important considerations whenevaluating patients with mild neurological deficits, who areeasy to underidentify and therefore less likely to receiverehabilitation.

Such documented limitations of the Berg have led manyclinicians to do more than one validated balance assessmentin order to identify deficits that may respond to treatment.Recently, a new and more comprehensive clinical balancetest, the Balance Evaluation Systems Test (BESTest), hasbeen developed that is essentially a battery of balance andmobility tests, borrowed from other validated tests such asthe Berg and Dynamic Gait Index. The BESTest was uniquelydesigned as a comprehensive clinical tool for evaluating sixdifferent balance control systems: biomechanical, stabilitylimits/verticality, anticipatory, reactive, sensory orientation,and stability in gait. Such system-specific assessment is help-ful in directing treatment and to ensure that a meaningfuldeficit is not overlooked. The BESTest has good interraterreliability [10] and good validity in discerning fallers fromnonfallers in patients with PD [11].

The BESTest, though comprehensive, valid, and reliable,is lengthy to administer and may not always be prac-tical in a busy clinical setting. Thus, a shorter versionof the BESTest, the Mini-BESTest, was developed using


psychometric techniques to reduce item redundancy andsimplify scoring [12]. This shorter version has excellentinterrater (ICC ≥ 0.91), and test-retest (ICC ≥ 0.88)reliability and similar in length to the Berg [13]. However itis currently unknown how the Mini-BESTest compares withthe Berg in detecting balance deficits in the PD population.

The purpose of this study was to explore the usefulnessof the Mini-BESTest compared to the Berg in evaluatingbalance in people with PD of varying severity. Specifically,we evaluated (1) the distribution of patients scores to look forceiling effects, (2) concurrent validity with severity of disease,and (3) the sensitivity/specificity of separating people who door do not have postural response deficits.

2. Methods

Ninety-seven participants with idiopathic PD participatedin the study. These participants were part of either alarger clinical study examining prospective fall risk or anexercise efficacy study. Therefore, the group here representsa convenience sample of participants with PD, and thedata for this paper was taken from their baseline visits.Inclusion criteria: all people in the study were diagnosed withidiopathic PD by a movement disorders neurologist. Peoplewere excluded from the study if they presented with cognitiveimpairment, prior orthopedic injuries, or impairments thatcould interfere with mobility such as artificial joints orperipheral neuropathy or prior brain surgery such as apallidotomy or deep brain stimulation. All participantssigned informed consent forms approved by the OregonHealth & Science University Institutional Review Board. Allwork was conducted in accordance with the declaration ofHelsinki (1964).

All participants came in for an assessment of theirbalance and mobility which included both clinical andinstrumented testing. The data presented in this paper istaken from the clinical scales: the Unified Parkinson’s DiseaseRating Scale (UPDRS) III Motor section, Hoehn & Yahr(H&Y) disease severity classification, and the Berg andthe Mini-BESTest. The testing was performed in the sameorder for each participant, and rest breaks were given asneeded to avoid fatigue. Other balance and gait assessmentsconducted during testing that were not included in thisanalysis included gait and sway analysis using wearableinertial sensors. Testing was conducted at the Oregon ClinicalTranslational Research Institute at Oregon Health & ScienceUniversity. All participants took their PD medication asnormally indicated and were tested in the ON state. Allof the participants except for two were currently takingsome form of PD medication. The testing was administeredby a trained examiner, overseen by a physical therapist.Participant characteristics are outlined in Table 1.

2.1. Clinical Tests

2.1.1. Mini-BESTest. The Mini-BESTest test is a 14-item testthat focuses on dynamic balance, specifically anticipatorytransitions, postural responses, sensory orientation, and

Table 1: Participant characteristics.

Variables Mean SD Range

UPDRS III 31.6 11.2 12–60

Hoehn & Yahr 2.3 0.6 1–4

Age (yr) 65.6 7.1 47–83

Time since dx (yr) 6.5 5.0 0–23

Height (cm) 172.6 9.5 152–198

Weight (kg) 79.2 15.6 43–120

Gender Male 59 Female 38

dynamic gait [12]. Each item is scored from (0–2); a score of0 indicates that a person is unable to perform the task while ascore of 2 is normal. The best score is the maximum amountof points, being 28.

2.1.2. Berg Balance Scale (Berg) [2]. The Berg is a 14-item testdesigned to measure the balance of older adults by assessingtheir performance of specific functional tasks [14]. Each taskis scored from (0–4), for a maximum of 56 points. The testindicates that a score of 41–56 is associated with a low fallrisk, 21–40 with a medium fall risk, and 0–20 with a high fallrisk [14].

2.1.3. Unified Parkinson’s Disease Rating Scale (UPDRS).Disease severity was evaluated using the UPDRS III motorcomponent [15]. This test has a maximum score of 108; eachitem is scored from 0-not affected through 4-most severelyaffected.

2.1.4. Hoehn and Yahr (H&Y). Postural response deficitswere identified as patients scoring 3 to 4 in the H&Y scale.[16]. A score of 3 and above indicates postural instability asdefined by an abnormal stepping response to a backwardspull on the shoulders. The H&Y scale is the most commonlyused method for evaluating the severity of PD [17], and thescale ranges from 0 (no symptoms of PD) to 5 (wheelchairbound).

2.2. Statistics. The STATA statistical package was used forboth calculations and graphics [18]. We describe the Bergand Mini-BESTest data for the 97 participants, using his-tograms and a scatter plot displaying the association betweenthe two variables. We used the bootstrap method to assessa P value for the skewness [19]. We also carried out aregression of UPDRS jointly on the two scores for the Bergand Mini-BESTest. This regression provided information onthe relative contributions of the Berg and Mini-BESTest forpredicting the UPDRS, each adjusted for the other measureusing added variable or partial correlation plots that showthe extent of information in each test that is not conveyedby the other test [20]. Finally, we considered the relativeperformance of the Berg and Mini-BESTest in terms ofreceiver operating characteristic (ROC) curves for classifyingpeople into two groups based on a threshold for the H&Yscore, to discriminate between mild PD (H&Y 1-2) versusmore severe PD (H&Y 3-4) [21].


0

5

10

15Fr

equ

ency

0 10 20 30 40 50 60

Berg balance score

(a)

0

5

10

15

Freq

uen

cy

0 5 10 15 20 25 30

Mini-BESTest score

(b)

0

10

20

30

40

50

60

Ber

gba

lan

cesc

ore

0 5 10 15 20 25 30Mini-BESTest score

(c)

Figure 1: Distribution of scores for the Berg Balance scale (a) and Mini-BESTest test (b), along with a scatter plot showing their relationshipto one another (c) for 97 patients with Parkinson’s disease.

3. Results

3.1. Distribution of Scores and Relation between Berg andMini-BESTest. The distribution of scores among the 97 par-ticipants with PD on the Mini-BESTest differed significantlyfrom the Berg (Figure 1). The Mini-BESTest scores weresignificantly less skewed than the Berg (Berg skewness= −2.3versus Mini-BESTest skewness = 0.93; P < 0.001). Usingthe bootstrap method, we found that, sampling from a pop-ulation with the shape of the Mini-BESTest histogram, thechance would be less than 0.001 of obtaining a skewness asextreme as that seen for Berg. The scatter plot in Figure 1(c)shows the relationship between the two measures.

The Mini-BESTest and Berg correlate significantly(r = 0.79; P < 0.001). However, people scoring the highestvalues in the Berg (i.e., 52–56; those with scores in theclinically accepted range as “normal”) had scores represent-ing approximately half of its maximum range in the Mini-BESTest. This suggests that the Mini-BESTest “spreads out”the compression (ceiling effect) at the top end of the Berg.

3.2. Relationship to PD Severity. Both the Mini-BESTest andBerg were moderately correlated with disease severity as

measured by the UPDRS. Figures 2(a) and 2(b) displaythe individual regression lines, indicating that the Berg andthe Mini-BESTest each have a significant correlation to theUPDRS (−0.39 and −0.51, P > 0.001, respectively).

Using a multiple regression of the UPDRS on both theMini-BESTest and the Berg, we determined how much eithertest compliments the other in the prediction of diseaseseverity. For linear regression prediction of the UPRDS, theBerg did not provide statistically significant informationin addition to the Mini-BESTest (t = 0.18; P = 0.86). Incontrast, the Mini-BESTest provided significant informationin addition to the Berg (t = −3.7; P = 0.001) to predictseverity of disease. The added variable plot in Figure 2(c)shows the extent of information in the Mini-BESTest forpredicting UPDRS, beyond that provided by the Berg. Thiswas significant (P < 0.001). The added variable plot inFigure 2(d) shows the extent of information in the Berg forpredicting UPDRS, beyond that provided by Mini-BESTest.This was not statistically significant (P = 0.86).

3.3. Identifying Mild Deficits. We compared the ability ofthe Berg and Mini-BESTest to differentiate PD patientswith and without clinical balance deficits. Participants with


0

10

20

30

40

50

60U

PD

RS

mot

orsc

ore

0 10 20 30 40 50 60

Berg balance score

(a)

0

10

20

30

40

50

60

UP

DR

Sm

otor

scor

e

0 5 10 15 20 25 30

Mini-BESTest score

(b)

−20

−10

0

10

20

30

−5 0 5

Mini-BESTest unexplained by Berg

UP

DR

Su

nex

plai

ned

byB

erg

(c)

−20

−10

0

10

20

30

−15 −10 −5 0 5 10

Berg unexplained by Mini-BESTest

UP

DR

Su

nex

plai

ned

byM

ini-

BE

STes

t

(d)

Figure 2: Scatter plots showing the relationship of the UPDRS motor score to (a) Berg Balance scale, and (b) Mini-BESTest score. Lowerpanels show (c) added value of the Mini-BESTest over Berg and, (d) added value of Berg over Mini-BESTest for predicting UPDRS motorscore.

and without clinical balance deficits were classified usingH&Y: H&Y 1-2 and H&Y 3-4. A score of H&Y 3 and 4identifies people with abnormal postural stepping responseto the backwards pull test or observable postural instability.Though the mean H&Y score was 2.3, the range was 1–4.Roughly one third (31 of 97) of the participants had a H&Y of3 or above, indicating postural instability as defined by H&Y.Figures 3(a) and 3(b) compare the distributions of Berg andMini-BESTest scores for people with H&Y 1-2 versus H&Y 3-4. ROC analysis was done to test the discriminative ability ofthese different balance tests to differentiate those people withand without abnormal postural responses.

The area under the ROC curves (AUC) differed for thetests; the AUC for the Berg = 0.84 ± 0.04 and the AUC forthe Mini-BESTest = 0.91 ± 0.03. The 2-sided P-value fortesting equality of the two AUC values was 0.05. A suggestedcut-off point for the Mini-BESTest to differentiate thosewith and without postural response deficits is > 21, yielding(sensitivity, specificity) = (89%, 81%). The nearest point tothis for the Berg is ≥52, yielding (Sensitivity, Specificity)= (77%, 74%). The points corresponding to these cut-offpoints are indicated by circles in Figure 3(c).

3.4. Most Difficult Items for People with PD. Individual itemsfrom both the Berg and the Mini-BESTest were ranked inorder of difficulty for the whole population of people withPD within this study and classified as “difficult” if a personhad a score less than perfect on that item (2 = perfect; 1 =some difficulty, or 0 = cannot perform) (Table 2). We foundthat 72% (10 out of 14) items on the Mini-BESTest presentedsome difficulty to at least one-third of the group versus only36% (5 out of 14 items) in the Berg.

4. Discussion

The results from this study suggest that the Mini-BESTestmay be more useful than the Berg in evaluating balance dis-orders in patients with PD, especially in those with mild PDor more subtle balance deficits. Specifically, results showedthat (1) although the Mini-BESTest had a high correlationwith the Berg, it did not have the same ceiling effects; (2)both the Berg and Mini-BESTest correlated with PD severitybut the Mini-BESTest added value to the Berg score; (3) theMini-BESTest test had better sensitivity/specificity then theBerg to identify people with abnormal postural responses.


Table 2: The Berg and Mini-BESTest individual items ranked from most difficult to least based on the % of participants with PD who didnot have normal scores. Difficulty with the test was determined if the participant did not receive a perfect score.

Berg test item Percentage (% with difficulty) Mini-BESTest item Percentage (% with difficulty) System (Mini-BEST)

Turning to look behind 70.1 Rise to toes 86.6 Anticipatory

Standing with one footin front

42.3 Single leg 81.4 Anticipatory

Reaching forward withoutstretched arms

40.2 TUG w/Cog 54.6 Gait

Standing on one foot 39.2 Pivot turn 51.5 Gait

Turn 360 degrees 30.9 Eyes Closed/foam 46.4 Sensory

Placing alternate foot onstool

27.8 Obstacle during Gait 46.4 Gait

Standing to sitting 11.3 Turn head with gait 41.2 Gait

Retrieving object fromthe floor

9.3 Incline eyes closed 33 Sensory

Sitting to standing 5.2 Backwards recovery 29.9 Postural

Standing with feettogether

4.1 Lateral recovery 29.9 Postural

Transfers 4.1 Change pace gait 13.4 Gait

Standing with eyesclosed

3.1 Forward recovery 13.4 Postural

Standing unsupported 3.1 Sit to stand 6.2 Anticipatory

Sitting unsupported 0 Eyes open stance 2.1 Sensory

The high correlation of the Mini-BESTest with the Bergsupports concurrent validity since the Berg remains oneof the most commonly used clinical scales for balanceassessment in people with PD. But importantly, we foundvery different test score distributions across patients withvaried levels of severity. Though neither test had a normaldistribution, the Mini-BESTest was significantly less skewed,indicating that there are less ceiling effects as has beenshown previously with the Berg [22]. These results are notsurprising since the Berg was originally intended for frailelderly and remains an excellent measure of balance deficitsfor those with more severe PD. The high sensitivity of theMini-BEST is important for clinicians who see patients withmild balance deficits who are seeking to identify and treatpotentially preventable mobility problems early in the diseaseprogression.

The Berg has been shown to have excellent test-retestreliability [3] and to correlate significantly with diseaseseverity in PD [23], and our results support the relationshipwith the UPDRS. Both exercise and physical therapy havebeen shown to improve UPDRS scores. Therapists needmeasures that reflect improvements with intervention socomparing the Mini-BESTest with the UPDRS establishesconcurrent validity of the new test with an established one.The novel information obtained from our study is thatwhile both the Berg and Mini-BESTest correlate with diseaseseverity, the Mini-BESTest adds value not included in theBerg, but the Berg does not add value to the Mini-BESTest.These findings suggest that the Mini-BESTest distinguishesamong PD subjects who all get similar, high scores in theBerg, and this information can add to the prediction ofdisease severity. A previous study demonstrated the Berg to

be useful in identifying balance impairments in people withvery severe PD (i.e., H&Y 4), but it could not discriminatesubgroups of H&Y scores successfully [24]. Here, we foundsimilar results in that the Mini-BESTest was more successfulthan the Berg at discriminating subgroups of PD severity asmeasured by the H&Y scale. Franchignoni et al. examinedthe clinimetric properties of the Berg with 57 participantswith PD [9]. They found excellent internal consistency,good correlations to other scales of disease severity, andquality of life, all agreeing with previously published work[4]. However, they did find, using a Rasch analysis, thatsome rating categories were not used and others wereunderutilized. The authors suggested that improving therating scale structure would improve the test. The sametype of Rasch analysis was performed on the full BESTest toobtain the shortened Mini-BESTest that excludes redundantor underused items [12].

The cut-off point of the Mini-BESTest for identifyingpatients with PD who had problems with the “Pull test” (i.e.,H&Y score of at least 3) was a score of 21. It is interesting thata similar cut-off point for the Mini-BESTest for identifyingpatients with PD who fall was a score of 20 [13]. Boththe Mini-BESTest and the Berg were sensitive (89% and77%, respectively) and specific (81% and 74%, respectively)in differentiating those with and without postural responsedeficits. Similarly, the Mini-BESTest was also shown to besensitive (88%) and specific (78%) in identifying PD patientswith a history of falls [13].

It has been suggested that postural instability in PDis multifactorial, therefore, a multitude of tests should beadministered by physical therapists [25, 26]. For example,the Berg does not include tests of postural reactions or


0

10

20

30

40

50

60B

erg

bala

nce

scor

e

H&Y 1-2 H&Y 3-4

(a)

0

5

10

15

20

25

30

H&Y 1-2 H&Y 3-4

Min

i-B

EST

est

scor

e

(b)

0

0.25

0.5

0.75

1

Sen

siti

vity

0 0.25 0.5 0.75 1

1-specificity

Berg balance scoreMini-BESTest score

(c)

Figure 3: (a) and (b) show the distribution of scores on the Berg Balance Scale (a) and the Mini-BESTest (b) when patients were separatedinto those without postural response deficits (H&Y 1-2) and those with postural deficits (H&Y 3-4). (c) compares the receiver operatorcharacteristics of the Berg Balance Scale and the Mini-BESTest to differentiate people with and without postural response deficits as measuresby H&Y classification.

dynamic gait, and, therefore, some deficits may be missed.Since the Mini-BESTest is essentially a combination of tests,this may be a reason it successfully identified people withmild balance deficits. As outlined in Table 2, each test itemprimarily tests one of 4 categories of balance: anticipatory,dynamic gait, reactive control, and sensory orientation.The Berg was not designed with such systems in mindbut if a system categorization is assigned to each item,the Berg items primarily evaluate anticipatory and sensorycontributions to balance. There are two additional systemsthat the Mini-BESTest evaluates, dynamic gait, and reactivepostural control, this may explain the added variable plotbeing significant for the Mini-BESTest adding value to theBerg in relating to disease severity. In other words, the Mini-BESTest usefully distinguishes among those persons that areoverly range compressed in the Berg. If a clinician is using theBerg for their PD patients, it may be beneficial to augmenttesting with the Dynamic Gait Index and the Pull test fromthe UPDRS. Dynamic gait (cognitive task with gait) andreactive postural control (response to perturbation) items

were the most difficult items for people with PD, balancesystems that are not assessed using the Berg.

Clinicians commonly use single-limb stance for balanceassessment. An example of a difference between testing itemsin the Berg and Mini-BESTest is the assessment of the single-limb stance (item #14 Berg, item #3 Mini-BESTest). In theBerg, the participant chooses either leg, and it is only this sidethat is assessed. Comparatively, the Mini-BESTest assessesboth the left and right leg and records the worst side. In thisstudy, when the Berg was used, assessing only one leg, 39%of the participants had some observable difficulty. When theMini-BESTest was used, assessing both left and right leg, 81%of the participants had some difficulty. Therefore, cliniciansshould test standing balance on both sides.

This study was limited to people with PD so it needs to berepeated in patients with other pathologies affecting balancecontrol. One potential limitation is that the order of testingwas not randomized so fatigue may have factored into testperformance. However, participants were given frequent restbreaks to avoid fatigue.


In conclusion, the Mini-BESTest is a novel, useful, andeasy to administer tool for balance assessment. Althoughthe Mini-BESTest had a high correlation with the Berg, itdid not have the same ceiling effects. Furthermore, boththe Berg and Mini-BESTest correlated with PD severitybut the Mini-BESTest added value to the Berg score inpredicting disease severity. Finally, the Mini-BESTest testhad better sensitivity/specificity than the Berg to identifypeople with abnormal postural responses. Taken together,these findings suggest that the Mini-BESTest is a promisingtool for discerning balance deficits in patients with mild tosevere PD.

Acknowledgments

This research is supported by National Institute of Aging(R37AG006457), National Institute of Neurologic Disordersand Stroke (RC1NS068678), the Kinetics Foundation, andOregon Clinical & Translational Institute (OCTRI) (ULIRR024140).

References


[2] K. O. Berg, S. L. Wood-Dauphinee, J. I. Williams, and B.Maki, “Measuring balance in the elderly: validation of aninstrument,” Canadian Journal of Public Health, vol. 83,supplement 2, pp. S7–S11, 1992.

[3] T. Steffen and M. Seney, “Test-retest reliability and minimaldetectable change on balance and ambulation tests, the 36-Item Short-Form Health Survey, and the Unified ParkinsonDisease Rating Scale in people with parkinsonism,” PhysicalTherapy, vol. 88, no. 6, pp. 733–746, 2008.

[4] A. A. Qutubuddin, P. O. Pegg, D. X. Cifu, R. Brown, S.McNamee, and W. Carne, “Validating the Berg Balance Scalefor patients with Parkinson’s disease: a key to rehabilitationevaluation,” Archives of Physical Medicine and Rehabilitation,vol. 86, no. 4, pp. 789–792, 2005.

[5] A. H. Newstead, M. R. Hinman, and J. A. Tomberlin,“Reliability of the Berg Balance Scale and balance master limitsof stability tests for individuals with brain injury,” Journal ofNeurologic Physical Therapy, vol. 29, no. 1, pp. 18–23, 2005.

[6] R. A. Geiger, J. B. Allen, J. O’Keefe, and R. R. Hicks, “Balanceand mobility following stroke: effects of physical therapy inter-ventions with and without biofeedback/forceplate training,”Physical Therapy, vol. 81, no. 4, pp. 995–1005, 2001.

[7] L. D. B. Thorbahn, R. A. Newton, and J. Chandler, “Use of theBerg balance test to predict falls in elderly persons,” PhysicalTherapy, vol. 76, no. 6, pp. 576–585, 1996.

[8] D. L. Kornetti, S. L. Fritz, Y. P. Chiu, K. E. Light, and C.A. Velozo, “Rating scale analysis of the Berg balance scale,”Archives of Physical Medicine and Rehabilitation, vol. 85, no.7, pp. 1128–1135, 2004.

[9] F. Franchignoni, C. A. Velozo, A. A. Qutubuddin, D. X. Cifu,and W. Carne, “Use of the Berg Balance Scale in rehabilitationevaluation of patients with Parkinson’s disease,” Archives ofPhysical Medicine and Rehabilitation, vol. 86, no. 11, pp. 2225–2226, 2005.

[10] F. B. Horak, D. M. Wrisley, and J. Frank, “The balanceevaluation systems test (BESTest) to differentiate balancedeficits,” Physical Therapy, vol. 89, no. 5, pp. 484–498, 2009.

[11] A. L. Leddy, B. E. Crowner, and G. M. Earhart, “Functionalgait assessment and balance evaluation system test: reliability,validity, sensitivity, and specificity for identifying individualswith parkinson disease who fall,” Physical Therapy, vol. 91, no.1, pp. 102–113, 2011.

[12] F. Franchignoni, F. Horak, M. Godi, A. Nardone, and A.Giordano, “Using psychometric techniques to improve thebalance evaluation systems test: the mini-bestest,” Journal ofRehabilitation Medicine, vol. 42, no. 4, pp. 323–331, 2010.

[13] A. L. Leddy, B. E. Crowner, and G. M. Earhart, “Utility ofthe mini-BESTest, BESTest, and BESTest sections for balanceassessments in individuals with Parkinson disease,” Journal ofNeurologic Physical Therapy, vol. 35, no. 2, pp. 90–97, 2011.

[14] K. Berg, S. Wood-Dauphinee, and J. I. Williams, “The balancescale: reliability assessment with elderly residents and patientswith an acute stroke,” Scandinavian Journal of RehabilitationMedicine, vol. 27, no. 1, pp. 27–36, 1995.

[15] S. Fahn and R. Elton, Unified Parkinson’s Disease Rating Scale,Macmillan Healthcare Information, Florham Park, NJ, USA,1987.


[17] C. Ramaker, J. Marinus, A. M. Stiggelbout, and B. J. van Hilten,“Systematic evaluation of rating scales for impairment anddisability in Parkinson’s disease,” Movement Disorders, vol. 17,no. 5, pp. 867–876, 2002.

[18] Stata Statistical Software, Release 11, StataCorp LP, CollegeStation, Tex, USA, 2009.

[19] A.C. Davison and D. Hinkley, Bootstrap Methods and TheirApplication, Cambridge Series in Statistical and ProbabilisticMathematics, Cambridge University Press, Cambridge, UK,8th edition, 2006.

[20] J. Neter, W. Wasserman, and M. Kunter, Applied LinearStatistical Models, Irwin, 1990.

[21] E. R. DeLong, D. M. DeLong, and D. L. Clarke-Pearson,“Comparing the areas under two or more correlated receiveroperating characteristic curves: a nonparametric approach,”Biometrics, vol. 44, no. 3, pp. 837–845, 1988.

[22] L. Blum and N. Korner-Bitensky, “Usefulness of the BergBalance Scale in stroke rehabilitation: a systematic review,”Physical Therapy, vol. 88, no. 5, pp. 559–566, 2008.

[23] K. J. Brusse, S. Zimdars, K. R. Zalewski, and T. M. Steffen,“Testing functional performance in people with Parkinsondisease,” Physical Therapy, vol. 85, no. 2, pp. 134–141, 2005.

[24] H. Tanji, A. L. Gruber-Baldini, K. E. Anderson et al., “Acomparative study of physical performance measures inParkinson’s disease,” Movement Disorders, vol. 23, no. 13, pp.1897–1905, 2008.

[25] J. V. Jacobs, F. B. Horak, V. K. Tran, and J. G. Nutt, “Multiplebalance tests improve the assessment of postural stabilityin subjects with Parkinson’s disease,” Journal of Neurology,Neurosurgery and Psychiatry, vol. 77, no. 3, pp. 322–326, 2006.

[26] L. E. Dibble and M. Lange, “Predicting falls in individualswith Parkinson disease: a reconsideration of clinical balancemeasures,” Journal of Neurologic Physical Therapy, vol. 30, no.2, pp. 60–67, 2006.


Research Article

The PIT: SToPP Trial—A Feasibility RandomisedControlled Trial of Home-Based Physiotherapy forPeople with Parkinson’s Disease Using Video-BasedMeasures to Preserve Assessor Blinding

Emma Stack,1, 2 Helen Roberts,1 and Ann Ashburn3

1 Faculty of Medicine, Academic Geriatric Medicine, University of Southampton, Southampton SO17 1BJ, UK2 Southampton General Hospital, Mailpoint 886, Tremona Road, Southampton, Hampshire SO16 6YD, UK3 Faculty of Health Sciences, University of Southampton, Southampton SO17 1BJ, UK

Correspondence should be addressed to Emma Stack, [email protected]

Received 25 July 2011; Accepted 22 August 2011


Copyright © 2012 Emma Stack et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Purpose. To trial four-week’s physiotherapy targeting chair transfers for people with Parkinson’s disease (PwPD) and explore thefeasibility of reliance on remote outcome measurement to preserve blinding. Scope. We recruited 47 PwPD and randomised 24 toa focused home physiotherapy programme (exercise, movement strategies, and cueing) and 23 to a control group. We evaluatedtransfers (plus mobility, balance, posture, and quality of life) before and after treatment and at followup (weeks 0, 4, 8, and12) from video produced by, and questionnaires distributed by, treating physiotherapists. Participants fed back via end-of-studyquestionnaires. Thirty-five participants (74%) completed the trial. Excluding dropouts, 20% of questionnaire data and 9% ofvideo data were missing or unusable; we had to evaluate balance in situ. We noted trends to improvement in transfers, mobility,and balance in the physiotherapy group not noted in the control group. Participant feedback was largely positive and assessorblinding was maintained in every case. Conclusions. Intense, focused physiotherapy at home appears acceptable and likely to bringpositive change in those who can participate. Remote outcome measurement was successful; questionnaire followup and furthertraining in video production would reduce missing data. We advocate a fully powered trial, designed to minimise dropouts andpreserve assessor blinding, to evaluate this intervention.

1. Introduction

Chair transfers, a common cause of falls [1, 2], are a keydomain of physiotherapy for people with Parkinson’s disease(PwPD) [3–5]. While weak lower limbs and inflexible,unstable trunks extend rising time [6–10], exercise shortenssit-to-stand times and PwPD can relearn motor sequences,facilitating movement through cueing [3, 4, 11–14].

In their 2007 evidence-based analysis of physical therapyin Parkinson’s disease (PD), Keus et al. [3] found supportiveevidence for improving the performance of transfers amongPwPD in just two studies. The potential to improve transfersamong PwPD has been underresearched since Kamsma et al.[11] and Nieuwboer et al. [12] evaluated the use of cognitive

movement strategies, the former in a randomised controlledtrial (RCT; n = 38), the latter in a nonrandomisedcontrolled trial (n = 33). Over 12 months, Kamsma etal.’s experimental group (mean age 68 years) practiceda seven-step sequence for safe rising (positioning hands,positioning feet, shifting to the seat edge, repositioninghands, leaning forward, rising into standing, and adoptingupright posture), a “logical structure” that offered “max-imum opportunity for controlled execution without timeconstraints.” Participants were found able to learn anddemonstrate the strategies and they reported their use in real-life situations, though the effects of training were activity-specific. Nieuwboer et al.’s participants (mean age 66 years)undertook six weeks of home-based physiotherapy aimed


at reducing specific difficulties during functional activitiesincluding rising which was based on cueing, consciouscontrol of movement, biomechanical compensation, andrepetition of movement in differing circumstances. The chairrising strategy taught entailed repositioning “the centre ofmass in relation to the base of support to compensate forslow trunk flexion and insufficient horizontal momentum.”Chair transfers improved significantly after treatment, whenmeasured against the Parkinson’s Activity Scale (PAS) [13].More recently, Mak and Hui-Chan [14], recruited 60 PwPDto an RCT comparing rising times after four week’s audio-visual cued task-specific sit-to-stand training, four week’sconventional mobility and strengthening exercise (for thetrunk and lower limbs, followed by sit-to-stand training),and no treatment. Rising times shortened after treatment inboth the cued group (mean age 63 years) and the exercisegroup (mean age 66 years), by 25% and 10%, respectively.

In 2006, the National Institute for Health and ClinicalExcellence called for further trials to investigate physiother-apy in PD [15]; however, trials overly focused on measuringcost-effectiveness and quality of life might overlook mean-ingful changes in performance that are the realistic targetsof physiotherapy, potentially reducing the chances of PwPDaccessing appropriate therapies. Blinding assessors is “one ofthe methodological safeguards” that ensures a trial’s “internalvalidity” [16]: inadvertent “unblinding” is an importantissue in rehabilitation research. In physiotherapy trials, anassessor’s blinding is jeopardized when a participant saysor does something that hints at, or confirms, their groupallocation: as face-to-face or telephone contact betweenassessor and participant is highly likely to break blinding, weinvestigated remote evaluation of video- and questionnaire-based measures by a blinded assessor.

In a feasibility RCT, we investigated whether focusedphysiotherapy (cueing, movement strategies, and exercise),increased transfer independence and speed while reducingdifficulty and brought secondary changes in gait, balance,posture, and quality of life. Key issues were the programme’spotential for a full-sized trial (we did not power this studyto test the significance of differences between groups orover time) and the feasibility and acceptability of methods(including the acceptability of participating in researchat home). In keeping with recent physiotherapy studiesinvolving PwPD [17, 18], we opted for a home-based trial,firstly to avoid participant travel being a barrier to anyone’sparticipation and secondly, as specialists advocate the deliv-ery of physiotherapy at home [4, 12] where “activities areproving problematic” [19].

2. Methods

We recruited PwPD from a clinic and support groups withinHampshire who

(1) had a working diagnosis of PD, stages I to IV [20],fulfilling the UK PDS Brain Bank diagnostic criteria[21]. The staging allowed us to compare grosslythe spectrum of PD in the intervention and controlgroups:

(i) stage I indicating mild unilateral symptoms,

(ii) stage II, bilateral symptoms without balanceimpairment,

(iii) stage III, postural instability but independentlymobile,

(iv) stage IV, severe PD although able to stand andwalk with assistance;

(2) self-reported chair transfers as

(a) being excessively slow and/or,

(b) requiring much effort, assistance, or repeatedattempts and/or,

(c) associated with a previous fall;

(3) scored at least 8/12 on The Middlesex Elderly Assess-ment of Mental State [22], a gross screen for cognitiveimpairment that we have used in previous studiesto identify anyone with PD at risk of being unableto give fully informed consent [2, 23] and that isone of the most commonly used assessments byoccupational therapists of mental state in PwPD [24];

(4) were willing and able to undertake all aspects of theintervention;

(5) were willing and able to complete the outcome mea-sures (albeit with help from another person in com-pleting questionnaires, if handwriting was problem-atic).

Southampton and South West Hampshire Ethics Com-mittee approved the project and all participants gave writteninformed consent to all aspects of the study, includingspecifically video recording. After responding to an initialinvitation to take part, interested parties were visited athome by one of three treating physiotherapists. Havinggiven their consent, a therapist completed the participant’sbaseline (week 0) assessment, after which the participant wasrandomised to either the intervention group (receiving phys-iotherapy) or the control group (receiving none) throughconcealed allocation.

2.1. Intervention. The physiotherapy group undertook afour-week-long, evidence-based [3, 11–14], home physio-therapy programme focused on chair transfers, comprising(1) supervised exercise (to enhance hip and knee extensorstrength and trunk stability and flexibility), (2) teaching andlearning movement strategies for safer and easier standingand sitting, and (3) verbal cueing. The intervention wasnot novel in content but in its intense, focused delivery.Respecting individuality and professionalism, the protocoldictated only that the physiotherapists provided no morethan 12 hours of input focused only on chair transfers (amaximum of one hour, three times per week for four weeks)including only portable equipment (like ankle weights); thetherapists decided if, when, and how intensively to useexercise, strategies, and/or cueing based on each partici-pant’s assessment and their clinical experience. The primaryobjective was to improve the ability to transfer, in terms of


independence, speed (timed sit-to-stand), and/or difficulty.Improvements in mobility, balance, posture, or quality of lifewould be secondary benefits.

2.2. Assessment. We assessed participants four times, at thesame time of day and point during an “on” phase, beforeand after intervention (weeks 0 and 4) and during followup(weeks 8 and 12). The physiotherapists’ video-recorded fivephysical performance tests for the blinded assessor toevaluate: PAS; sit-to-stand [14]; Standing Start 180 DegreeTurn Test (SS-180) [25]; functional reach (FR) [26]; theUnified PD Rating Scale Posture Item [27]. All five tests havebeen used previously with PwPD (and three were developedspecifically for PwPD), and we followed the publishedprotocols without modification. A single camera, stopwatch,and metre-rule were used across the study but, as a home-based study, the chairs that participant’s rose from during thePAS varied (but each person used their same chair at everyassessment point).

The physiotherapists’ also left the participants two ques-tionnaires to complete at home and return in a stampedaddressed envelope: PD Self-Assessed Disability Scale (SAS)[28] and 15D instrument of health-related quality of life(HR-QOL) [29]. On SAS, participants indicated on a five-point scale how much effort 25 everyday tasks took: bestscore 25, worst score 125. On HR-QOL, they indicated ona five-point scale the state of 15 aspects of health: best score15, worst score 75.

The physiotherapists transferred edited clips of eachperformance (but nothing superfluous) onto DVD for oneindependent assessor (blind to group allocation) to evaluate,alongside returned questionnaires and feedback. Indepen-dence and effort demonstrated during the PAS Chair Transfersection was rated from zero to eight (worst to best score).Sit-to-stand was stopwatch-timed (during the PAS) from thepoint when a participant started to move until they attainedstable standing. The SS-180 mean turn time was calculatedfrom two turns (one in each direction) to walk towards atarget. FR was the mean of three maximal forward reachesin standing. Posture was rated as the degree of stoop notedwhen participant’s stood for the FR, rated zero to four (bestto worst score). Sound was turned off during evaluation toprevent inadvertent unblinding but restored afterwards on asample to quality check that the therapists had used standardinstructions during data collection. During piloting, wefound it unacceptable to evaluate FR from video, as thenumbers on the metre-rule were (1) too small to read in thewide image necessary to rate posture and quality-check testconduct and (2) obscured by the reaching arm. So during theRCT physiotherapists rated FR in situ.

In their final week of involvement, we gave participantsan anonymous feedback form (including space to comment)posing the following questions about the acceptability ofhome-based research, the randomisation, intervention, andassessments.

Did you find it difficult to fit taking part in the studyinto your routine?

Was it difficult to find enough space for video record-ing at home?

How do you feel about not being in the group thathad physiotherapy?

Did you find physiotherapy helpful (and why)?

Do you think you had too little, enough, or too muchphysiotherapy (and why)?

Did you feel comfortable with video recording?

Were the tests quick enough to complete?

Were the questionnaires boring or difficult to com-plete or difficult to post back?

3. Results

We recruited 47 PwPD (median age 74 years; median yearssince diagnosis seven), including 45 who found transfersexcessively slow, 39 who found transfers an excessive effort,and 17 who had fallen transferring; 13 participants (28%)reported all three indicators. Participant median Hoehnand Yahr stage was III and half the participants had fallenrepeatedly in the previous year. The control group (n = 23)and treatment group (n = 24) were similar in characteristicsand baseline performances (Table 1).

Thirty-five participants (74%) completed: four (9%)dropped out by week 4, five (11%) by week 8, and 12 (26%)by week 12. Of the final 12 drop-outs, eight (67%) werefrom the treatment group, through illness. Dropouts wererepresentative of the whole sample age (median 73 years) andyears since diagnosis (median 8) but 11/12 (92%) were atHoehn and Yahr stage III or IV (in comparison with 79%of the whole sample).

Over the intervention period (weeks 0 to 4), as outlinedin Table 2, the physiotherapy group median PAS score tendedto improve (from 4 to 6) while that of the controls tendedto worsen (from 6 to 4). The tendency to improvementin the physiotherapy group continued throughout followup(median score reaching 7 by week 12) while the controlmedian returned to baseline. Median sit-to-stand timestended to shorten from weeks 0 to 4 (by 14%, from 2.2 s to1.9 s, in both groups) and to continue shortening to week12 (reaching 1.5 s in the physiotherapy group and 1.7 s inthe controls, reductions of 32% and 23% from baseline,resp.). The physiotherapy group median SAS score tendedto improve slightly by week 4 (by one point, from 50 to 49)while that of the controls tended to worsen (by 7 points,from 52 to 59); median scores tended to worsen in bothgroups by week 12 (the physiotherapy group deterioratingfrom baseline by 2 points, the controls by 12 points).

Over the intervention period, the physiotherapy groupmedian SS-180 time tended to shorten (by 17%, from 5.3 s to4.4 s) while that of the controls tended to lengthen (by 5%,from 3.7 s to 3.9 s). Throughout followup, the tendency toimprovement in the physiotherapy group continued (medianreaching 3.8 s by week 12, a reduction of 28% from baseline)while controls turn time remained longer than at baseline.The physiotherapy group median FR tended to improveslightly by week 4 (15%, from 19.2 cm to 22.0 cm) while that


Table 1: Participant characteristics at baseline (n = 47).

Variable Value Control group (n = 23) Treatment group (n = 24) All (n = 47)

Age (years) Median (IQR), range 74 (70–78), 58–86 75 (69–77), 64–82 74 (69–77), 58–86

Gender Men (n) 18 (78%) 17 (71%) 35 (74%)

Years since diagnosis Median (IQR), range 7 (4–12), 1–19 8 (4–11), 1–30 7 (4–12), 1–30

Hoehn and yahr (grade)

I 1 0 1 (2%)

II 5 4 9 (19%)

III 10 12 22 (47%)

IV 7 8 15 (32%)

UPDRS Median (IQR), range 30 (18–45), 9–52 26 (21–38), 10–60 28 (20–41), 9–60

12-month fall historyNo falls (n) 7 4 11 (23%)

Single fall (n) 6 6 12 (26%)

Repeated falls (n) 10 14 24 (51%)

Indication for physiotherapy

Transfers excessively slowly n (%) 23 (100) 22 (92) 45 (96)

Transfers a considerable effort n (%) 20 (87) 19 (79) 39 (83)

History of falls transferring n (%) 10 (43) 7 (29) 17 (36)

Primary outcome measures

PAS chair transfer score Median (IQR), range 5 (4–6), 0–8 4 (4–6), 2–8 4 (4–6), 0–8

Sit-to-stand time (s) Median (IQR), range 2.2 (1.6–3.7), 0.8–11.1 2.1 (1.5–3.2), 0.8–7.2 2.2 (1.5–3.2), 0.8–11.1

SAS score Median (IQR), range 54 (41–70), 37–104 50 (43–63), 36–90 51 (41–65), 36–104

Secondary outcome measures

SS-180 turn time (s) Median (IQR), range 3.8 (3.4–6.8), 1.8–45.6 5.5 (3.8–8.4), 2.2–43.5 5.3 (3.5–7.4), 1.8–45.6

FR (cm) Median (IQR), range 21 (17–25), 10–33 18 (16–21), 9–32 20 (16–23), 9–33

UPDRS posture score Median (IQR), range 1 (1-2), 0–4 1 (1-2), 0–3 1 (1-2), 0–4

HR-QOL score Median (IQR), range 30 (26–32), 21–44 30 (28–37), 16–47 30 (27–35), 16–47

IQR = interquartile range; range = minimum to maximum.

Table 2: Changes in outcomes (weeks 0–12, by group) in participants who completed a measure on at least three occasions; values presentedare medians (interquartile range).

Primary outcomes Group Week 0 Week 4 Week 8 Week 12

PAS chair transfer (score)Control (n = 18) 6 (4–7) 4 (4–6) 6 (4–7) 6 (3–7)

Physiotherapy (n = 19) 4 (4–6) 6 (4–7) 7 (5–8) 7 (4–8)

Sit-to-stand time (s)Control (n = 20) 2.2 (1.4–3.2) 1.9 (1.3–2.8) 2.0 (1.4–2.2) 1.7 (1.3–2.2)

Physiotherapy (n = 18) 2.2 (1.6–3.1) 1.9 (1.4–2.0) 1.7 (1.0–2.4) 1.5 (1.2–2.0)

SAS (score)Control (n = 18) 52 (40–64) 59 (45–71) 60 (48–66) 64 (50–77)

Physiotherapy (n = 16) 50 (43–59) 49 (43–67) 58 (50–66) 52 (43–60)

Secondary outcomes

SS-180 turn time (s)Control (n = 16) 3.7 (3.1–6.8) 3.9 (3.1–7.0) 4.1 (2.7–9.6) 3.9 (2.8–5.9)

Physiotherapy (n = 16) 5.3 (3.9–6.7) 4.4 (3.4–6.4) 3.7 (3.4–4.8) 3.8 (3.1–6.1)

FR (cm)Control (n = 16) 20.9 (15.7–25.2) 21.0 (15.0–24.3) 21.7 (13.6–25.8) 19.7 (17.4–27.7)

Physiotherapy (n = 13) 19.2 (17.5–21.9) 22.0 (20.0–25.0) 22.8 (20.3–25.8) 25.5 (19.6–30.2)

UPDRS posture (score)Control (n = 19) 1 (1-2) 1 (1-2) 1 (1-2) 1 (1-2)

Physiotherapy (n = 19) 1 (1-2) 1 (1-2) 1 (1-2) 1 (1-2)

HR-QOL scoreControl (n = 14) 29 (26–31) 30 (25–33) 29 (25–33) 31 (24–34)

Physiotherapy (n = 14) 29 (26–36) 30 (28–36) 32 (28–38) 29 (27–34)

of the controls changed minimally (0.5%, from 20.9 cm to21.0 cm); the tendency to improvement in the physiotherapygroup continued (median reaching 25.5 cm by week 12, a33% increase from baseline) while the control’s worsened (tomedian 19.7 cm, a 6% decrease from baseline). We detected

little change in posture or quality of life in either group. Onthe UPDRS posture item, both groups were rated a median1 (IQR 1-2) at every assessment point. The median HR-QOLscore of both groups worsened by one point (from 29 to 30)by week 4; at week 12 the physiotherapy group median had


returned to baseline while that of the control group was twopoints worse.

Thirty-nine participants (83%) returned anonymousfeedback forms at the end of the study, 20 from thephysiotherapy group and 19 controls. Few participants founddifficulty fitting in the study or finding space for the video-based assessments (3/35 in each case). Of 17 controls,seven expressed disappointment about not being in thetreatment group (“I would have liked to see if it could helpmy condition”), while ten were ambivalent, for example,“someone had to be in this group and I was one of them.” Two ofthe treatment group (10%) found the intervention unhelpfulbut 18 (90%) found it helpful, reporting they had

(1) learned new skills or exercises (“Despite long-termParkinson’s I still learned new ways”),

(2) found movement or exercise easier (“Easy to followinstruction, made easier by being in my home; I putthem into practice during my daily round; they havebecome part of my routine”),

(3) gained useful advice and support from a goodtherapist (“Guidance managing my disability”).

Eight (40%) felt they had insufficient therapy time, as onehour was an inadequate representation (“She never saw me atmy worst”) or they needed encouragement (“Physiotherapyprompts me into being regular with my own efforts”) and“the more the better” (“It works well at the time andmore would have been an advantage”). But twelve (60%)suggested it fitted their routine (“I was not overburdened”),they were tired afterwards (“Quite tired by the end”), andmore might have been repetitive (“I did not feel bored ordisinterested”). Most felt comfortable being video-recorded(35/36) and found assessment acceptably quick (33/35). Fewfound the questionnaires difficult to complete (2/35) orreturn (1/39) or found them boring (2/36). Participantswrote few comments; although overwhelmingly positive, oneparticipant felt “filming was unrepresentative; she never sawme totally unable to move or having to crawl on the floor.”

Reliance on video and questionnaires preserved assessorblinding fully: the assessor learned no-one’s group allocationas there were no distinguishing features in 600-plus silentvideo clips.

Of a potential 1316 measurements (seven outcomesmeasured for 47 participants four times), 131 (10%) weremissing as participants had dropped-out (Table 3); of theremaining measurements, 185/1185 (16%) were missing orunusable. FR was the only test someone declined to attempt,and a documentation error (later rectified) invalidated 14%of potential FR data. Of the tests evaluated from video, theSS-180 had most missing/unusable data, 15% after excludingdropouts: records were discounted if the protocol had beenfollowed incorrectly (e.g., the participant turned in the samedirection on both trials), if editing invalidated the clip (e.g.,ending before the turn was complete) or if the recording wasinadequately lit. Timed sit-to-stand, PAS, and posture scorehad percentages of unusable data below 10%, after excludingdropouts.

4. Discussion

This study revealed a trend to improved transfers, mobility,and balance among PwPD after physiotherapy. It would befeasible to deliver this focused programme quickly and easilyin the home. Over a quarter of participants had multipledifficulties with transfer speed, effort, and stability, all ofwhich physiotherapists can address, yet Keus et al. [30]found transfers to be a physiotherapy priority in just 14%of cases, behind gait (74%), posture (49%), and balance(37%). While controls deteriorated, our intervention grouptended to continue to improve over followup (sit-to-standtime decreased 14% by week 4 and 32% by week 12), whichsuggests continual refinement of newly learned strategies. Asthe need is evident and intervention possible, as suggested bythese results and others [1–14], transfer training (preferablyat home, where people can deploy strategies learned) shouldbe integral to physiotherapy for PwPD. Illness, fatigue, anda lack of perceived benefit are commonly reasons whyPwPD discontinue exercise regimes [31]. Illness among ourtreatment group accounted for most dropouts (as was thecase when Nieuwboer et al. [12] lost 15% of their recruits toa home-based physiotherapy trial over 12 weeks), and someparticipants reported extreme fatigue after physiotherapy:this intervention warrants selective application.

The importance of physiotherapists’ expertise and rela-tionships with patients in the quality and outcome of treat-ment is well recognised [30, 32]. From feedback, our sampleappeared pro-physiotherapy (they learned new skills andways of managing their condition) and pro-research (theyunderstood their role within the study). Although severalcontrols were disappointed, none dropped out following ran-domisation. Experienced clinicians in physiotherapy trialsenhance the participant experience and data collection.

Among our video-based evaluations, drop-outs account-ed for two-thirds of the missing data; technical difficultieswith video production errors (such as filming in too darka setting or in too confined or cluttered a space or editingthe clip prepared for the blinded assessor too harshly) andprotocol/documentation errors (such as having the personperforming the SS-180 turn twice in the same rather thanthe opposite direction or recording the FR as habituallydone during clinical practice rather than in the standardway used in the study) accounted for the rest. The layout,light levels, and contents of an individual’s home are theirchoice and researchers cannot expect the ideal conditionsfor data collection that they might expect in a purpose-built movement laboratory; although they can prepare thearea used for video-recording to a certain extent, it is notreasonable to impose whole-scale modifications. It can bedifficult to know at the time of recording that a clip will betoo poorly illuminated for the assessor when they see it later.Similarly, a participant may move out of the camera’s scopein an unexpected way that only comes to light when editingthe clip. The more complex the test, the more likely it is thata proportion will be obscured or lost. On balance, we believethat losing a proportion of data recorded in the home isoutweighed by the benefits of the inclusivity that home-basedresearch offers to the people who wish to participate. Specific


Table 3: Reasons for missing and unusable data, by measure type.

Measure typeNumber of potentialmeasurements

Number of data missing Number of data unusable

Participant drop-outTest declined, omitted, or notrecorded or questionnaire notreturned

Error in testing or recording,poor-quality recording, orquestionnaire incomplete

In situ—real time

FR 188 19 (10%) 18 (10%) 27 (14%)

Video based

PAS 188 19 (10%) 9 (5%) 4 (2%)

Transfer time 188 19 (10%) 9 (5%) 2 (1%)

SS-180 time 188 19 (10%) 10 (5%) 18 (10%)

Posture 188 19 (10%) 9 (5%) 5 (3%)

Postal questionnaires

SAS 188 18 (10%) 24 (13%) 9 (5%)

HR-QOL 188 18 (10%) 25 (13%) 16 (9%)

Total 1316 131 (10%) 104 (8%) 81 (6%)

training and monitoring of therapists would improve theseaspects of data collection. This feasibility study revealed thatwhile experienced clinicians make good researchers, they areunlikely to have been trained in all the necessary researchskills beforehand and they are also likely to have developedways of conducting and recording tests in their professionalpractice that differ from the study protocols. Again, onbalance, we believe that the benefit of employing experiencedclinicians outweighs the costs involved in employing andtraining them in research skills which may be entirely new.

FR was the only test declined for fear of falling (by oneindividual on three occasions) and was the only assessmentnot feasible in the home, in this study. After dropouts, themajor causes of missing/unusable questionnaire data wereunreturned questionnaires (13% were missing) and incom-plete questionnaires (omitted answers invalidated 9% ofotherwise complete HR-QOL questionnaires). Feedback didnot indicate problems but prompts or collection (both withcost and ethical implications) would have increased returns.Haapaniemi et al. [29], who received 15% of their HR-QOL questionnaires incomplete, used regression analysis topredict missing data, an option we would advocate.

Our assessment battery, which took approximately 20minutes to complete in the home, measured the difficulty,slowness, and dependence associated with transfers. Speedimproved by week 4 in both our groups, associated withan improvement in PAS score in the treatment group anddeterioration among the controls. Others have demonstratedchanges in transfer strategy after training while speedremains unchanged [33] and have stressed that functionand stability are more important than speed [4]. We couldrecommend both the PAS score and sit-to-stand time asprimary outcome measures in future, as the demand onparticipants is low and evaluation from video was associatedwith relatively little missing data. In light of the numbers ofunreturned and incomplete questionnaires, and its breadth,we would not advocate the SAS in a similar way.

While it may be impossible to blind a trial participant asto whether they have actively taken part in a physiotherapyintervention [34], it should be possible to keep the assessorblind. Even when it is possible to blind the assessor, this is notalways done or reported. In the present study, using silentvideo and questionnaires preserved total assessor blindingand was associated with other advantages over face-to-face ortelephone contact. Working without distraction from editedclips reduced the time (and money) spent travelling to,and engaging with, participants by the experienced assessor.Video facilitates reliability testing, team review, illustrationof findings, and quality control. Participants found theassessments acceptable, and (with the exception of FR)the assessments were feasible using standard equipment.Evaluation of video by a blinded assessor has been usedsuccessfully in fields such as gastric surgery [35] and paincontrol [36]; here we have demonstrated its feasibility inmovement analysis. Video should only be used with explicitconsent ensuring precautions are taken to (a) avoid recordinganyone else’s image or conversation and (b) maintain thesubject’s safety [37]: a tripod-mounted, battery-poweredcamera is safer than one which occupies the recorder’s handsand presents a trip hazard.

We make the following recommendations for a futureRCT. Employ experienced clinicians, teach them the requiredresearch skills, and monitor fidelity to the recording pro-tocol throughout the trial. When calculating a sample size,consider that our 19 controls scored a mean 5.11 (1.97)on the PAS Chair Transfer at baseline, which decreased amean 0.32 (1.29) to a mean 4.79 (1.47) at week 4; theintervention group (n = 20) scored a mean 4.55 (1.39)at baseline, which increased a mean 0.90 (1.80) to a mean5.45 (1.61) at week 4. Over-recruit by 25% to compensatefor dropouts. Offer controls some intervention, in light ofthe multiple transfer difficulties of PwPD and our controls’feedback. Use more sensitive measures of posture (such astragus-to-wall distance) and quality of life than we trialled.


As specifically supervised exercise was a component of theintervention in this trial, we did not ask participants toreport any unsupervised exercise they may have undertaken;we suggest that researcher’s consider recording the latter ina full RCT. While the performances evaluated in this studywere analysed by one blinded assessor, they were recordedby the unblinded trial physiotherapists: while the recordingsallowed us to quality check test conduct, employing anindependent recorder to collect and prepare clips for theassessor might reduce potential bias even further.

5. Conclusion

If the measures are suitable, an intense, focused physiother-apy programme delivered at home by experienced physio-therapists is likely to bring about positive change in thosewho can participate. We recommend a fully powered trial(over-recruiting to offset dropouts and offering the controlsan incentive to remain) using remote outcome measurement(especially silent video assessment), as far as possible, topreserve blinding: missing/unusable data can be reducedby comprehensive training in video recording/editing andfollowing up/collecting questionnaires or replacing missinganswers using statistical methods. The key implication of thisfeasibility trial is that a sample of PwPD older than in theearlier studies discussed, appeared to derive benefit from aprogramme that was shorter than most of the earlier studiesand one that practicing clinicians could roll out withoutadditional training or equipment.

Funding

This work (Promoting Independent Transfers: The South-ampton Trial of Physiotherapy for Parkinson’s Disease, PIT:SToPP) was supported by the Parkinson’s UK (Reference G-0507).

Acknowledgments

The authors wish to acknowledge the support of DeNDRoN(South Coast) and the Parkinson’s UK branches of Hamp-shire and the contribution of Chartered PhysiotherapistsElizabeth Ashdown, Kate Gahr, and Martina Kaucka.

References

[1] A. Ashburn, E. Stack, C. Ballinger, L. Fazakarley, and C. Fitton,“The circumstances of falls among people with Parkinson’sdisease and the use of Falls Diaries to facilitate reporting,’’Disability and Rehabilitation, vol. 30, no. 16, pp. 1205–1212,2008.



[4] A. Ashburn, D. Jones, R. Plant et al., “Physiotherapy for peoplewith PD in the UK: an exploration of practice,” InternationalJournal of Therapy and Rehabilitation, vol. 11, no. 4, pp. 160–167, 2004.

[5] W. J. Mutch, A. Strudwick, S. K. Roy, and A. W. Downie,“Parkinson’s disease: disability, review, and management,”British Medical Journal, vol. 293, no. 6548, pp. 675–677, 1986.

[6] V. K. Ramsey, T. A. Miszko, and M. Horvat, “Muscle activationand force production in Parkinson’s patients during sit tostand transfers,” Clinical Biomechanics, vol. 19, no. 4, pp. 377–384, 2004.

[7] M. K. Y. Mak, O. Levin, J. Mizrahi, and C. W. Y. Hui-Chan,“Joint torques during sit-to-stand in healthy subjects andpeople with Parkinson’s disease,” Clinical Biomechanics, vol.18, no. 3, pp. 197–206, 2003.

[8] L. M. Inkster, J. J. Eng, D. L. MacIntyre, and A. Jon Stoessl,“Leg muscle strength is reduced in Parkinson’s disease andrelates to the ability to rise from a chair,” Movement Disorders,vol. 18, no. 2, pp. 157–162, 2003.

[9] M. K. Y. Mak and C. W. Y. Hui-Chan, “Switching of movementdirection is central to Parkinsonian bradykinesia in sit-to-stand,” Movement Disorders, vol. 17, no. 6, pp. 1188–1195,2002.

[10] E. Nikfekr, K. Kerr, S. Attfield, and D. E. Playford, “Trunkmovement in Parkinson’s disease during rising from seatedposition,” Movement Disorders, vol. 17, no. 2, pp. 274–282,2002.

[11] Y. P. T. Kamsma, W. H. Brouwer, and J. P. W. F. Lakke,“Training of compensational strategies for impaired grossmotor skills in Parkinson’s disease,” Physiotherapy Theory andPractice, vol. 11, no. 4, pp. 209–229, 1995.

[12] A. Nieuwboer, W. de Weerdt, R. Dom, M. Truyen, L. Janssens,and Y. Kamsma, “The effect of a home physiotherapy programfor persons with Parkinson’s disease,” Journal of RehabilitationMedicine, vol. 33, no. 6, pp. 266–272, 2001.

[13] A. Nieuwboer, W. de Weerdt, R. Dom, K. Bogaerts, and G.Nuyens, “Development of an activity scale for individualswith advanced Parkinson disease: reliability and ’on-off ’variability,” Physical Therapy, vol. 80, no. 11, pp. 1087–1096,2000.


[15] National Institute for Health and Clinical Excellence (NICE),“Parkinson’s Disease: diagnosis and management in pri-mary and secondary care,” Clinical Guideline 35, 2006,www.nice.org.uk/CG035 .

[16] R. W. Poolman, P. A. Struijs, R. Krips et al., “Reporting ofoutcomes in orthopaedic randomized trials: does blinding ofoutcome assessors matter?” Journal of Bone and Joint Surgery,Series A, vol. 89, no. 3, pp. 550–558, 2007.

[17] J. Nocera, M. Horvat, and C. T. Ray, “Effects of home-based exercise on postural control and sensory organization inindividuals with Parkinson disease,” Parkinsonism and RelatedDisorders, vol. 15, no. 10, pp. 742–745, 2009.


[19] D. Jones and J. E. Playfer, “Parkinson’s disease,” in PhysicalManagement for Neurological Conditions, M. Stokes and E.Stack, Eds., chapter 6, Elsevier, 3rd edition, 2011.



[21] A. J. Hughes, S. E. Daniel, L. Kilford, and A. J. Lees, “Accuracyof clinical diagnosis of idiopathic Parkinson’s disease: aclinico-pathological study of 100 cases,” Journal of NeurologyNeurosurgery and Psychiatry, vol. 55, no. 3, pp. 181–184, 1992.

[22] A. Shiel and B. A. Wilson, “Performance of stroke patients onthe Middlesex Elderly Assessment of Mental State,” ClinicalRehabilitation, vol. 6, no. 4, pp. 283–289, 1992.

[23] E. Stack, K. Jupp, and A. Ashburn, “Developing methodsto evaluate how people with Parkinson’s Disease turn 180◦:an activity frequently associated with falls,” Disability andRehabilitation, vol. 26, no. 8, pp. 478–484, 2004.

[24] K. H. O. Deane, C. Ellis-Hill, K. Dekker, P. Davies, and C. E.Clarke, “A survey of current occupational therapy practice forParkinson’s disease in the United Kingdom,” British Journal ofOccupational Therapy, vol. 66, no. 5, pp. 193–200, 2003.

[25] E. Stack and A. Ashburn, “Early development of the Standing-start 180◦ Turn Test,” Physiotherapy, vol. 91, no. 1, pp. 6–13,2005.

[26] P. W. Duncan, D. K. Weiner, J. Chandler, and S. Studenski,“Functional reach: a new clinical measure of balance,” Journalsof Gerontology, vol. 45, no. 6, pp. M192–M197, 1990.

[27] E. Lang and S. Fahn, “Assessment of PD,” in Quantification ofNeurologic Deficit, T. Munsat, Ed., chapter 21, Butterworths,Methuen, Mass, USA, 1989.

[28] R. G. Brown, B. MacCarthy, M. Jahanshahi, and C. D.Marsden, “Accuracy of self-reported disability in patients withparkinsonism,” Archives of Neurology, vol. 46, no. 9, pp. 955–959, 1989.

[29] T. H. Haapaniemi, K. A. Sotaniemi, H. Sintonen, and E.Taimela, “The generic 15D instrument is valid and feasible formeasuring health related quality of life in Parkinson’s disease,”Journal of Neurology, Neurosurgery and Psychiatry, vol. 75, no.7, pp. 976–983, 2004.

[30] S. H. J. Keus, B. R. Bloem, D. Verbaan et al., “Physiotherapyin Parkinson’s disease: utilisation and patient satisfaction,”Journal of Neurology, vol. 251, no. 6, pp. 680–687, 2004.

[31] B. Caird, “Physiotherapy benefits patients with PD,” ClinicalRehabilitation, vol. 3, pp. 11–16, 1989.

[32] J. A. K. Moffett and P. H. Richardson, “The influence of thephysiotherapist-patient relationship on pain and disability,”Physiotherapy Theory and Practice, vol. 13, no. 1, pp. 89–96,1997.

[33] P. K. Schot, K. M. Knutzen, S. M. Poole, and L. A. Mrotek,“Sit-to-stand performance of older adults following strengthtraining,” Research Quarterly for Exercise and Sport, vol. 74, no.1, pp. 1–8, 2003.

[34] E. Stack, “Physiotherapy: the ultimate placebo,” PhysiotherapyResearch International, vol. 11, no. 3, pp. 127–128, 2006.

[35] S. Fiorucci, A. Mencarelli, A. Meneguzzi et al., “Co-administration of nitric oxide-aspirin (NCX-4016) andaspirin prevents platelet and monocyte activation and protectsagainst gastric damage induced by aspirin in humans,” Journalof the American College of Cardiology, vol. 44, no. 3, pp. 635–641, 2004.

[36] L. O’Brien, A. Taddio, M. Ipp, M. Goldbach, and G. Koren,“Topical 4% amethocaine gel reduces the pain of subcuta-neous measles-mumps-rubella vaccination,” Pediatrics, vol.114, no. 6, pp. e720–e724, 2004.

[37] D. Kunkel and E. Stack, “Falls and their Management,” inPhysical Management for Neurological Conditions, M. Stokesand E. Stack, Eds., chapter 20, Elsevier, 3rd edition, 2011.


Research Article

Gait Difficulty, Postural Instability, and Muscle Weakness AreAssociated with Fear of Falling in People with Parkinson’s Disease

Margaret K. Y. Mak,1 Marco Y. C. Pang,1 and Vincent Mok2

1 Department of Rehabilitation Sciences, The Hong Kong Polytechnic University, Hung Hom, Hong Kong2 Department of Medicine and Therapeutics, Prince of Wales Hospital, The Chinese University of Hong Kong, Hong Kong

Correspondence should be addressed to Margaret K. Y. Mak, [email protected]

Received 24 May 2011; Accepted 12 August 2011


Copyright © 2012 Margaret K. Y. Mak et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

The present study aimed to examine the contribution of gait impairment, postural stability and muscle weakness to the level offear of falling in people with Parkinson’s disease (PD). Fifty-seven community-dwelling individuals with PD completed the study.Fear of falling was assessed by the Activities-specific Balance Confidence (ABC) scale. Postural stability and gait difficulty weredetermined by the posture and gait subscores of the Unified Parkinson’s Disease Rating Scale (UPDRS-PG). A Cybex dynamometerwas used to measure isokinetic knee muscle strength. Individuals with PD achieved a mean ABC score of 73.6±19.3. In the multipleregression analysis, after accounting for basic demographics, fall history and disease severity, the UPDRS-PG score remainedindependently associated with the ABC score, accounting for 13.4% of the variance (P < 0.001). The addition of knee musclestrength significantly improved the prediction model and accounted for an additional 7.3% of the variance in the ABC score (P <0.05). This is the first study to demonstrate that the UPDRS-PG score and knee muscle strength are important and independentdeterminants of the level of fear of falling in individuals with PD. Improving balance, gait stability and knee muscle strength couldbe crucial in promoting balance confidence in the appropriately targeted PD population.

1. Introduction

Fear of falling (FoF) is a common and potentially seriousproblem in people with Parkinson’s disease (PD). Previousstudies have consistently reported that community-dwellingindividuals with PD have a greater FoF than age-matchedhealthy subjects [1–4]. The level of FoF is further increased inthose who have had a fall history [5]. In a prospective study,we found that FoF is also a significant risk factor for predict-ing future falls [4]. While some level of FoF has a protectiverole against falls, irrational FoF, either too much or too little,may increase fall risk. Delbaere et al. [6] have recently ad-dressed this complex psychological factor in a large cohortof older population and revealed that discrepancies betweenpsychological and physiological risk factors in those who hadexcessive or unduly low level of FoF. However, only thosewith excessive FoF had a higher risk of injurious falls. Re-peated falls may lead to avoidance of activity, physical de-conditioning, and increased institutionalization. Therefore,

interventions aiming to enhance balance confidence have thepotential to reduce fall risk in appropriately targeted individ-uals with PD.

To design effective treatment intervention, it is crucial tounderstand the factors that determine FoF. In people withPD, FoF was found to be associated with postural sway instanding and posture and gait impairment as measured bythe unified PD rating scale (UPDRS) [1], one-leg stancetime, timed-up-and-go time, 6-minute walk distance, andthe UPDRS motor score [7]. Jacobs et al. [2] reported thatthe combination of the pull test, the gait item of the UPDRS,and one-leg-stance time was better than single items inpredicting FoF. However, the regression model used in theirstudy did not include factors that could contribute to theprediction of FoF, such as demographic data, disease severity,and fall history. In addition, the association between musclestrength and FoF has not been examined. We recently foundthat recurrent PD fallers had more lower extremity muscleweakness than PD nonfallers and single fallers [5]. Deficits


in muscle power were found to associate with slower gaitvelocity and increase fall risk in individuals with PD [8].It is, therefore, important to determine the contribution ofmuscle strength in predicting FoF. The present study aimedto examine the factors that determine FoF in people with PD.Specifically, we examined balance and gait instability as wellas muscle strength, as these are significant fall risk factors inpeople with PD [3, 9].

2. Methods

A convenience sample of 57 individuals with PD completedthe study. PD participants were recruited from movementdisorders clinics in Hong Kong and the Hong Kong Parkin-son’s Disease Association, which is a patient self-help group.All patients were diagnosed by neurologists according tothe United Kingdom PD Society Brain Bank Criteria [10].All subjects were recruited on a volunteer basis. Informedconsent was obtained from each participant in accordancewith the 1964 Declaration of Helsinki, and all experimentalwork was carried out with the approval of the universityethics committee. To be included in this study, subjectswere required to be between 40 and 85 years of age, medi-cally stable, able to walk 6 metres at least three times withand without an assistive device, and able to understandsimple commands (minimental state examination score≥24 [11]). Subjects were excluded if they had neurologicalconditions other than idiopathic PD, exhibited posturalhypotension, visual disturbance, or vestibular dysfunctionaffecting balance, or had significant cardiovascular or mus-culoskeletal disorders limiting locomotion or balance. Allindividuals with PD were tested within 2 hours after medi-cation, that is, during the “on” phase of the medication cycle(Table 1).

3. Procedure

All evaluations were carried out at the Hong Kong Polytech-nic University gait and motion research laboratory. Demo-graphic data including age, body mass, height, and medi-cations were recorded. We measured disease severity by theHoehn and Yahr staging scale (HY) [12] and the motor com-ponent of the UPDRS [13, 14]. FoF was estimated by theactivities-specific balance confidence (ABC) scale [15]. Theknee muscle strength of participants was measured by aCybex Norm dynamometer. Information on the number offall events over the past 12 months was obtained by patientinterview. Participants were classified as fallers if they suf-fered at least one fall in the past 12 months. A fall is defined as“an event during which a subject comes to rest on the groundor at some lower level, not as the result of a major in-trinsic event for example, syncope, stroke and seizure, oroverwhelming hazard” [16].

Fear of falling was measured by the validated Chineseversion of the ABC scale [17]. Participants were asked torate their self-perceived balance confidence level from 0 (noconfidence at all) to 100 (full confidence) for completing 16activities of daily living. The mean score was calculated for

Table 1: Subject characteristics.

People with PD (N = 57)

Demographics

Age (years) 63.7 (8.5)

Height (cm) 161.2 (8.1)

Weight (kg) 61.1 (10.1)

Female gender, n (%) 22 (38.6)

Fallers, n (%) 19 (33.3)

Parkinson’s disease characteristics

Years since diagnosis of Parkinson’sdisease (years)

7.6 (4.6)

Hoehn and Yahr stage (0–5) 2.5 (1.0)#

UPDRS—motor score III (0–108) 22.6 (6.5)

UPDRS-PG (0–16) 4.0 (2.0)#

Knee muscle strength (Nm) 34.4 (13.3)

ABC score (0–100) 73.6 (19.3)

Data shown are means (standard deviations), #median (interquartile range),ABC: activities-specific balance confidence, UPDRS: unified Parkinson’sdisease rating scale, UPDRS-PG: unified Parkinson’s disease rating scale(items 27–30).

each subject, with a minimum score of 0 to a maximum of100. A lower ABC score indicates greater FoF.

The unified PD rating scale motor examination (UPDRS-III) is a valid tool used to assess the level of motor impair-ment and disability in individuals with PD [13, 14]. It con-sists of 14 items which assess PD-specific impairments. Eachitem scores from 0 to 4, with 0 indicating absence of impair-ment and 4 indicating severe impairment. In this study, thesum of items 27–30 (i.e., rising from a chair, posture, gait,and postural stability (UPDRS-PG) was used to documentthe postural instability and gait difficulty of PD participants[1].

Knee muscle strength was quantitatively assessed by a Cy-bex Norm isokinetic dynamometer (Lumex, Inc., Ronkonk-oma, NY, USA). The more affected lower extremity, whichwas determined by a higher unilateral UPDRS-III score, wasassessed. Participants were seated with their lower leg at 90◦

of knee flexion, and a strap and a footplate were attachedto their lower leg and feet, respectively. Participants werestabilized by trunk and thigh straps during the test. Theinvestigator then measured an anatomical zero when theknee was passively moved to full extension. Participants wereinstructed to perform isokinetic concentric and eccentriccontraction of the knee flexors and extensors from 10◦ to70◦ of flexion at an angular speed of 90◦/s. The order ofthe 4 testing conditions was randomized. Participants wereallowed to practice each type of contraction at their sub-maximal effort 2 times, which was followed by the test trialwhen 3 maximum concentric or eccentric contractions wereperformed. Participants were given a 3-minute rest betweeneach mode of contraction. The average value of the peaktorque (Nm) among the 3 test trials was obtained, and thesum of mean concentric and eccentric knee muscle strength


was used for further analysis. Overall, these strength testingprocedures lasted for 20 minutes.

4. Statistical Analysis

All statistical analyses were performed using SPSS 17.0,and a significance level of 0.05 (2-tailed) was set for allstatistical tests. The Shapiro Wilk statistic was used to checkdata normality. Descriptive analysis was performed for thedemographic data and variables of interest. Bivariate corre-lation analyses were performed. Pearson product momentcorrelation was performed to establish the relationshipbetween the ABC score and knee muscle strength, as thedata were normally distributed. For the UPDRS-PG score,which is an ordinal data, the relationship with the ABCscore was determined by Spearman’s rho. A hierarchicalmultiple linear regression model (enter strategy) was usedto determine the contribution of the UPDRS-PG score andknee muscle strength to the ABC score after accounting forother potential contributing factors (e.g., demographic data,fall history, and disease severity measured by the HY stagingscore). Age, duration of PD, fall history, and the HY stagingscores were first entered into the regression model followedby the UPDRS-PG scores and knee muscle strength.

5. Results

The mean ABC score for individuals with PD was 73.5±19.3.Individuals with PD had a median HY score of 2.5 ± 1.0,indicating mild-to-moderate disease severity. The medianUPDRS-PG score was 4.0 ± 2.0, implying mild gait andpostural instability. The mean knee muscle strength was34.4 ± 13.3 Nm. Correlation analysis showed that the ABCscore was positively correlated with knee muscle strength(r = 0.301, P = 0.029) and inversely correlated with theUPDRS-PG score (r = −0.661, P < 0.001). These findingsindicate that a higher level of FoF was associated with greaterknee muscle weakness and increased gait instability andpostural difficulty. The results of the regression model showthat after adjusting for basic demographics, fall history, anddisease severity, the UPDRS-PG score remained indepen-dently associated with the FoF level, accounting for 13.4%of the variance (Model 2, Table 2). The addition of kneemuscle strength significantly improved the model predictionby 7.3% (Model 3, Table 2). A total of 47.9% of the variancein the ABC score was predicted by the final regression model(F6,56 = 6.895, P < 0.001). Among all the variables, theUPDRS-PG score was the most important determinant of theABC score, as reflected by the magnitude of the regressioncoefficient (β = −0.531).

6. Discussion

Our PD participants had a mean ABC score of 73.6 ± 19.3,indicating that they had moderate level of FoF. This findingis consistent with the published data [1, 2, 4, 5, 7]. Thenegative association between FoF and the UPDRS-PG scoreconcurs with previous findings. Excessive FoF was shown to

be negatively correlated with the UPDRS-PG score [1, 7],centre of pressure sway during standing [1], and Berg’s bal-ance score, tandem Romberg, and timed up and go time [18]in individuals with PD. Our finding extends that reported byJacobs et al. [2] that postural instability and gait impairmentas measured by the UPDRS-PG score is an importantdeterminant of FoF, after accounting for demographic data,fall history, and disease severity in individuals with PD. TheUPDRS-PG score alone accounts for 13.4% of the variance ofthe ABC score. The UPDRS-PG score quantifies participants’standing upright posture, response to retropulsion, sit-to-stand transfer, and gait stability. Stooped posture in peoplewith PD was found to be destabilizing [19] and capableof predicting future falls [9]. In addition, people with PDare known to be slow and inflexible in response to externalperturbation, especially to a backward pull [20–22]. Walkingand rising from a chair have often been reported to be fall-related activities [23, 24]. For example, 24%–46% of indi-viduals with PD were reported to have fallen during walkingand turning and 15% of individuals with PD fell duringtransferring from sitting to standing [5, 23, 24]. Greaterpostural instability and gait difficulty in individuals withPD will lead to less perceived self-confidence in performingbalance activities, hence an increased level of FoF.

Previous studies reported that people with PD hadreduction in knee and ankle muscle strength [25–27], whichwas correlated with sit-to-stand performance [28] and gaitvelocity [26]. Our study is the first to report that knee musclestrength, which accounted for 7.3% of the variance of theABC score after accounting for demographic data and theUPDRS-PG score, is another important determinant of FoF.In a recent study, we reported that recurrent PD fallers hadsignificantly more reduced lower extremity muscle strengththan single fallers [5]. These recurrent fallers also perceivedthat “muscle gives way” was associated with their falls. Kneemuscle strength is crucial for maintaining stability in anupright position. Weakness in this muscle group could givethe patients the perception that their “muscles give way”while in the standing position and lead to a lack of confi-dence in performing standing or walking activities. Lowerextremity muscle strength was independently associated withreduced bone mass in individuals with PD [27]. Further-more, knee extensors muscle strength of the more affectedside was a significant fall predictor [9]. When combined withexcessive FoF, muscle weakness may restrict individuals’ ac-tivities, lead to further muscle weakness and accelerated lossof bone mass, and increase the risk of fall-related fracture.

To prevent falls in people with PD, treatment interven-tions should enhance both physical function and balanceconfidence. A recent systematic review reported that in olderadults, exercise was the most commonly used intervention toimprove balance confidence [29]. The exercise interventionsinclude strength, balance, and gait training. Combined cog-nitive behavioral education (i.e., identification of fall riskfactors, discussion of coping strategies for falling, and as-sertiveness training) and exercise training were found to beeffective in enhancing balance confidence and reducing therisk of falls in older people [30]. Based on the significantassociation between FoF and postural and gait impairment


Table 2: Multiple regression analysis for predicting the ABC score.

Independent variable R2 R2 change B (S.E.) β P value

Model 1 0.272 0.272

Age −0.827 (0.307) −0.359 0.010∗

Years since diagnosis −0.111 (0.553) −0.028 0.841

Fall history −5.982 (4.991) −0.159 0.237

HY stage −7.243 (5.606) −0.190 0.203

Model 2 0.406 0.134

Age −0.636 (0.287) −0.276 0.032∗

Years since diagnosis 0.029 (0.507) 0.007 0.954

Fall history −4.682 (4.576) −0.124 0.312

HY stage 0.645 (5.676) 0.017 0.910

UPDRS-PG −4.630 (1.439) −0.458 0.002∗∗

Model 3 0.479 0.073

Age −0.554 (0.274) −0.240 0.049∗

Years since diagnosis −0.113 (0.484) -0.028 0.816

Fall history −2.386 (4.427) −0.063 0.593

HY stage 2.828 (5.443) 0.074 0.606

UPDRS-PG −5.371 (1.394) −0.531 <0.001∗∗∗

Knee muscle strength 0.040 (0.016) 0.285 0.016∗∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, B: unstandardized regression coefficient, S.E.: standard error, β: standardized regression coefficient, ABC: activities-

specific balance confidence, UPDRS-PG: unified Parkinson’s disease rating scale (items 27–30), HY: Hoehn and Yahr.

and knee muscle weakness, clinicians may consider incorpo-rating muscle strengthening programmes, as well as improv-ing patients’ postural and gait stability, in their fall pre-vention programs. We believe that the promotion of balanceconfidence can prevent the vicious cycle of activity restric-tion, physical deconditioning, further decline in self-per-ceived balance confidence, and future falls. Further interven-tional study is needed to prove this postulation.

We acknowledge that our study has certain limitations.To be included in the study, participants needed to be able towalk freely to undertake the gait assessments. Our findings,therefore, are not generalisable to individuals with PD withsignificant gait impairments. Our assessments were alsorestricted to “on phase” periods. It is possible that conductingassessments during the “off” phase of treatment would in-crease their sensitivity. In addition, FoF is associated withmany factors. However, we could not include many predict-ing variables in the regression analysis due to our small sam-ple size. Our model was able to predict 47.9% of the varianceof the ABC score. Other physical factors such as freezingof gait and cognitive psychological factors such as cognitiveimpairment, anxiety, and depression could contribute tothe level of FoF. Finally, this is a cross-sectional study. Wecould not establish a causal relationship between posturalimpairment, gait difficulties, muscle weakness, and FoF.Further research should address the temporal relationshipbetween postural and gait impairment as well as muscleweakness and FoF.

To conclude, postural instability, gait difficulty, and kneemuscle weakness are important determinants of the level ofFoF. The clinical implication of our study is that the balance

confidence of people with PD may be enhanced through pro-moting muscle strength, balance, and gait stability, therebypreventing activity restriction and physical deconditioningand reducing fall risk. Further intervention study is neededto prove this postulation.

Acknowledgment

The study was supported by S. K. Yee Medical Foundation(ZH61).

References


[2] J. V. Jacobs, F. B. Horak, V. K. Tran, and J. G. Nutt, “Multiplebalance tests improve the assessment of postural stabilityin subjects with Parkinson’s disease,” Journal of Neurology,Neurosurgery and Psychiatry, vol. 77, no. 3, pp. 322–326, 2006.

[3] M. K. Y. Mak and M. Y. C. Pang, “Balance confidence andfunctional mobility are independently associated with falls inpeople with Parkinson’s disease,” Journal of Neurology, vol.256, no. 5, pp. 742–749, 2009.

[4] M. K. Y. Mak and M. Y. C. Pang, “Fear of falling is indepen-dently associated with recurrent falls in patients with Parkin-son’s disease: a 1-year prospective study,” Journal of Neurology,vol. 256, no. 10, pp. 1689–1695, 2009.

[5] M. K. Y. Mak and M. Y. C. Pang, “Parkinsonian single fallersversus recurrent fallers: different fall characteristics and clin-ical features,” Journal of Neurology, vol. 257, no. 9, pp. 1543–1551, 2010.


[6] K. Delbaere, J. C. T. Close, H. Brodaty, P. Sachdev, and S.R. Lord, “Determinants of disparities between perceived andphysiological risk of falling among elderly people: cohortstudy,” British Medical Journal, vol. 341, no. 7770, p. c4165,2010.

[7] C. A. Lohnes and G. M. Earhart, “External validation of abbre-viated versions of the activities-specific balance confidencescale in Parkinson’s disease,” Movement Disorders, vol. 25, no.4, pp. 485–489, 2010.

[8] N. E. Allen, C. Sherrington, C. G. Canning, and V. S. C.Fung, “Reduced muscle power is associated with slowerwalking velocity and falls in people with Parkinson’s disease,”Parkinsonism and Related Disorders, vol. 16, no. 4, pp. 261–264,2010.

[9] M. D. Latt, S. R. Lord, J. G. L. Morris, and V. S. C. Fung,“Clinical and physiological assessments for elucidating fallsrisk in Parkinson’s disease,” Movement Disorders, vol. 24, no.9, pp. 1280–1289, 2009.

[10] A. J. Hughes, S. E. Daniel, L. Kilford, and A. J. Lees, “Accuracyof clinical diagnosis of idiopathic Parkinson’s disease: a clin-ico-pathological study of 100 cases,” Journal of NeurologyNeurosurgery and Psychiatry, vol. 55, no. 3, pp. 181–184, 1992.

[11] M. F. Folstein, S. E. Folstein, and P. R. McHugh, “Mini-mentalstate: a practical method for grading the state of patients forthe clinician,” Journal of Psychiatric Research, vol. 12, no. 3, pp.189–198, 1975.


[13] S. Fahn and R. Elton, “Unified parkinson’s disease ratingscale,” in Recent Developments in Parkinson’s Disease, S. Fahn,C. D. Marsden, D.B. Caine, and M. Goldstein, Eds., vol. 2, pp.153–163, Macmillan Health Care Information, Florham Park,NJ, USA, 1987.

[14] S. Fahn and R. Elton, “Unified parkinson’s disease ratingscale,” in Recent Developments in Parkinson’s Disease, S. Fahn,C. D. Marsden, D.B. Caine, and M. Goldstein, Eds., vol. 2, pp.293–304, Macmillan Health Care Information, Florham Park,NJ, USA, 1987.

[15] L. E. Powell and A. M. Myers, “The Activities-specific BalanceConfidence (ABC) scale,” Journals of Gerontology Series A:Biological Sciences and Medical Sciences, vol. 50, no. 1, pp.M28–M34, 1995.

[16] M. E. Tinetti, M. Speechley, and S. F. Ginter, “Risk factorsfor falls among elderly persons living in the community,” NewEngland Journal of Medicine, vol. 319, no. 26, pp. 1701–1707,1988.

[17] M. K. Mak, A. L. Lau, F. S. Law, C. C. Cheung, and I. S.Wong, “Validation of the Chinese translated activities-specificbalance confidence scale,” Archives of Physical Medicine andRehabilitation, vol. 88, no. 4, pp. 496–503, 2007.

[18] F. Franchignoni, E. Martignoni, G. Ferriero, and C. Pasetti,“Balance and fear of falling in Parkinson’s disease,” Parkinson-ism and Related Disorders, vol. 11, no. 7, pp. 427–433, 2005.

[19] J. V. Jacobs, D. M. Dimitrova, J. G. Nutt, and F. B. Horak, “Canstooped posture explain multidirectional postural instabilityin patients with Parkinson’s disease?” Experimental BrainResearch, vol. 166, no. 1, pp. 78–88, 2005.


[21] D. Dimitrova, J. G. Nutt, and F. B. Horak, “Postural muscleresponses to multidirectional translations in patients with

Parkinson’s disease,” Journal of Neurophysiology, vol. 91, no. 1,pp. 489–501, 2004.

[22] J. V. Jacobs, F. B. Horak, K. van Tran, and J. G. Nutt, “An alter-native clinical postural stability test for patients with Parkin-son’s disease,” Journal of Neurology, vol. 253, no. 11, pp. 1404–1413, 2006.

[23] B. R. Bloem, Y. A. M. Grimbergen, M. Cramer, and A. H.Zwinderman, “Prospective assessment of falls in Parkinson’sdisease,” Journal of Neurology, vol. 248, no. 11, pp. 950–958,2001.



[26] M. Nallegowda, U. Singh, G. Handa et al., “Role of sensoryinput and muscle strength in maintenance of balance, gait, andposture in Parkinson’s disease,” American Journal of PhysicalMedicine and Rehabilitation, vol. 83, no. 12, pp. 898–908, 2004.

[27] M. Y. C. Pang and M. K. Y. Mak, “Muscle strength is signif-icantly associated with hip bone mineral density in womenwith Parkinson’s disease: a cross-sectional study,” Journal ofRehabilitation Medicine, vol. 41, no. 4, pp. 223–230, 2009.

[28] M. Paasuke, J. Ereline, H. Gapeyeva, K. Joost, K. Mottus, andP. Taba, “Leg-extension strength and chair-rise performance inelderlywomen with Parkinson’s disease,” Journal of Aging andPhysical Activity, vol. 12, pp. 511–524, 2004.

[29] C. J. Bula, S. Monod, C. Hoskovec, and S. Rochat, “Inter-ventions aiming at balance confidence improvement in olderadults: an updated review,” Gerontology, vol. 57, no. 3, pp. 276–286, 2011.

[30] S. Tennstedt, J. Howland, M. Lachman, E. Peterson, L. Kasten,and A. Jette, “A randomized, controlled trial of a group inter-vention to reduce fear of falling and associated activity restric-tion in older adults,” Journals of Gerontology Series B: Psycho-logical Sciences and Social Sciences, vol. 53, no. 6, pp. P384–P392, 1998.


Clinical Study

A Manipulation of Visual Feedback during Gait Training inParkinson’s Disease

Quincy J. Almeida and Haseel Bhatt

Sunlife Financial Movement Disorders Research and Rehabilitation Centre, Wilfrid Laurier University, Waterloo,ON, Canada N2L 3C5

Correspondence should be addressed to Quincy J. Almeida, [email protected]

Received 23 June 2011; Accepted 5 July 2011


Copyright © 2012 Q. J. Almeida and H. Bhatt. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

Visual cues are known to improve gait in Parkinson’s disease (PD); however, the contribution of optic flow continues to bedisputed. This study manipulated transverse line cues during two gait training interventions (6 weeks). PD subjects (N = 42)were assigned to one of three groups: treadmill (TG), overground (OG), or control group (CG). Participants walked acrosslines placed on either treadmills or 16-meter carpets, respectively. The treadmill (TG) offered a reduced dynamic flow from theenvironment, while lines presented on the ground (OG) emphasized optic flow related to the participant’s own displacement. Bothinterventions significantly improved (and maintained through retention period) step length, thus improving walking velocity. Onlythe OG improved in the TUG test, while only the TG showed hints of improving (and maintaining) motor symptoms. Since gaitimprovements were found in both training groups, we conclude that by reducing optic flow, gait benefits associated with visualcueing training can still be achieved.

1. Introduction

Individuals with Parkinson’s disease (PD) have been shownto walk with a stooped posture, limited arm swing, slowvelocity, and small shuffling steps that can often lead to falls[1]. Sensory cueing strategies such as auditory, tactile, andvisual cues have often been used to help walking in PD.Stein and Glickstein [2] suggested that of all these modalities,visual cues are most effective in improving PD gait. It is notclear, however, whether improvements might be the resultof improved use of optic flow, greater attention directedtowards walking, or cortically driven planning of discretesteps that bypass the basal ganglia.

Optic flow is a prominent theory that is often put forwardto explain the benefits associated with using transverse lines.This theory suggests that transverse lines improve walkingdue to the stripes accentuating the flow of the surroundingenvironment as one moves through space [3, 4]. This notionof optic flow has been strongly supported by Azulay et al. [5]that believe the lines emphasized optic flow which improvedgait velocity and stride length in PD participants. Optic

flow has been previously manipulated through either virtualreality or a projected tunnel screen [6, 7], and in each case,manipulation was presented by changing the surroundingenvironment. An interesting method of manipulating visualinformation from the surrounding environment is to havepeople walk on a treadmill. Biomechanically, the differencesthat exist between treadmill and overground walking arenegligible [8]. Interestingly, however, walking on a treadmillallows a reduction of typical optic flow that would normallybe associated with every day walking. Song and Hidler [8]and Frankel-Toledo et al. [9] acknowledge that subjects ona treadmill do not receive the same optic flow as they dowhen walking overground. Bello et al. [10] proposed thatgait improvements in PD treadmill walking are caused bythe subject’s ability to strategically use the distance fromthe front of the treadmill as a static visual cue. Contrarily,a study by Azulay et al. [5] used stroboscopic lighting tosuppress optic flow by transforming stripes on the floorto static cues, resulting rather in a deterioration of gaitin PD patients. This contradicting evidence indicates thatlittle is known as to how much, if any, optical flow is


needed to improve gait in PD. Thus, comparing overgroundand treadmill training with identical visual cues providesa unique opportunity to evaluate how optic flow mightcontribute to gait improvements.

Fundamental to these gait deficits is the inability toproduce a normalized step length [11]. Many popularvisually guided cues have been shown to improve step lengthincluding the inverted walking stick, projected laser beam[12, 13], and parallel lines [14]. It has been well establishedthat transverse lines an inch wide or more have been bestshown to facilitate locomotion [15]. Jiang and Norman [16]found that transverse lines assisted in the initiation of gaitin PD individuals. However, most studies that implementtransverse lines have often only conducted single sessions[16–18]. Morris et al. [17] showed that a single cueing sessionwas effective in regulating stride length in PD and that atraining effect emerged leading to improvements two hoursafter visual cues were removed. However, the potential forlong-term cue training to lead to even longer lasting benefitsto gait has yet to be studied. Interestingly, the only casestudy (with an n = 1) using transverse lines as a long-termcueing intervention revealed potential benefits [19]. Thus,more research must explore transverse lines as a long-termcueing therapy for Parkinson’s disease.

Unfortunately, most of the above studies failed to admin-ister a retention assessment; hence, any persisting long-termimprovements to gait have yet to be determined. Also, anassessment of gait transference to a more functional testsuch as the timed up and go (TUG) has not been used, aswell as potential symptomatic improvements (UPDRS motorscores). Through the administration of these tests, we canachieve greater insight into the underlying mechanism ofimprovement with the use of transverse line cues during gait.

One method of manipulating the provision of transverselines is to modify the context in which they are provided fortraining. For example, integrating transverse line cues on atreadmill is novel, as it provides step cues but within a morestatic background. In contrast, transverse lines provided overthe length of a carpet would move past any individual relativeto the rest of the surrounding environment. Thus, our studycompared two different methods of providing transverseline cues: (1) traditional overground gait training (withtransverse lines) and (2) treadmill training. Our primaryoutcome measure was step length, while additional measuresincluded UPDRS motor scores, lower limb strength gains,TUG times, and other spatiotemporal aspects of gait. Allvariables were assessed at baseline (pretest), after a 6 weekrehabilitation phase (posttest), and 6 weeks later (retentiontest).

2. Methods

2.1. Subjects. The study included a total of 42 participantsthat were assigned to one of three PD groups: treadmill(TG), overground (OG), or control (CG). All participants(recruited through a database from the Sun Life MovementDisorders Research and Rehabilitation Centre, Wilfrid Lau-rier University, Canada) were diagnosed with Parkinson’s

disease and then randomized and matched for overall, aswell as PD specific demographics (based on a prescreeningassessment).

Each participant tested was confirmed to have clinicallytypical PD from at least one movement disorders neurolo-gist. All PD patients were responsive to anti-Parkinsonianmedication and were in an optimally medicated or “on”medication state at the time of all training and testingsessions.

Participants were excluded from the study if they hada past history of neurological conditions other than PD ororthopaedic or visual disturbances that severely impairedwalking ability. Also, participants were removed if they wereunable to independently walk down an 8 meter GAITRitecarpet for a total of 10 trials. Each participant was informedof the requirements of the study and signed institutionallyapproved consent forms, according to the declaration ofHelsinki (BMJ 1991; 302: 1194).

2.2. Materials. Data was collected in two different rooms, agymnasium and a laboratory measuring approximately 20 m× 10 m and 9.5 m× 6 m, respectively. Gait data was collectedin the gymnasium on a GAITRite carpet (GAITRite, CIRSystem, Inc., Clifton, NJ, USA) which measured 8 m long× 0.92 m wide and contained sensors that provided footfallinformation to an attached computer. The 30-second chairstand and TUG test were conducted in the laboratory.Materials needed for the two tests included a straight backchair, a taped line 3 meters away from the chair, and a stopwatch. Two Biodex Gait Trainer 2 treadmills were used forthe treadmill group, and three 16-meter black landscapingcarpets were used for the overground group. Transverse lineswere created using white athletic tape.

2.3. Protocol

2.3.1. UPDRS Severity Score. All participants’ motor symp-toms were assessed by a blinded movement disorders special-ist using the UPDRS Section 3.

2.3.2. Timed Up and Go (TUG). TUG test required partici-pants to sit in a chair and when told “go”, participants wereasked “to stand up, walk to the taped line, turn around,and sit back in the chair as quickly and safely as possible.”Two trials were performed and time was recorded using astopwatch. The purpose of the TUG was to assess functionalmobility of PD participants and track gait changes over time[20].

2.3.3. 30-Second Chair Stand. The 30-second chair standrequired all participants to be seated in a chair and whentold “go” to rise to a full stand position and sit back downagain. This was repeated as many times as possible in aspan of 30 seconds. Two trials were performed, and the totalnumber of stands was recorded. This measure was used toidentify any lower limb strength gains that may result fromthe intervention.


2.3.4. GAITRite Walking. All participants were requested towalk down an 8-meter GAITRite carpet “at a normal casualwalking speed” for a total of 5 trials. If participants neededfurther explanation, they were asked to walk down the carpetas though they were “walking down the street.” Participantsstarted 1 meter before the carpet and told to walk 2 stepsbeyond the end of the carpet to ensure gait initiation andtermination were not processed in data collection. Footfallinformation was collected to an attached computer, and thefollowing gait measures were obtained: gait velocity (cm/s),cadence (steps/min), mean step length (cm), double supporttime (s), step time (s), step-to-step variability, step-timevariability, and double support variability.

2.3.5. Training Protocol. Participants completed gait training3 times a week for 6 weeks (18 sessions in total). Each gaitsession spanned 30 minutes with a mandatory 2-minutebreak every 8 minutes. However, participants were allowedadditional rest if necessary but were required to walk atotal of 24 minutes for the gait session to be consideredcomplete. All participants were “on” medication at the timeof pre-, post-, and retention testing and during training.All training sessions were conducted at the same scheduledtime. Spotters were provided for all participants to ensuresafety. In both training groups, visual cues were provided(on ground or treadmill) with the use of white lines (seedescription below). To standardize the step length requiredduring training, we selected a separation between lines thatwas a minimum of 8% greater than the initial step length ofany of the groups. Thus, based on previous research [12] andalso this 8% requirement, the white lines were separated by70 cm. This ensured that from one consecutive heel strike tothe next, participants in both the overground and treadmillgroup trained with an equivalent distance between cuesteps. Furthermore, in order to control for training velocity,stepping was monitored using a timer over the distancecovered for the overground group, while velocity could beset manually for the treadmill group. In both cases, trainingvelocity was based on each individuals predetermined self-paced velocity.

(a) Overground Group. Overground gait training requiredparticipants to walk down equally spaced transverse lines,presented on a 16-meter carpet. The cues were white lines oftape equally distributed at a standardized length on the blackbackground carpet. Participants trained at the same walkingspeed that was measured at pretest (GAITRite analysis). Thiswas achieved by requiring participants to completely clearthe carpet within a specified amount of time. Participantswere asked to walk across the lines, turn, and continueback. A spotter would also assist in tracking time to ensureparticipants completed the trial in the allotted time.

(b) Treadmill Group. Treadmill gait training required par-ticipants to walk on a treadmill presented with equally dis-tributed standardized transverse white lines. All participantswalked at the speed determined at pre-test. This speed was

Table 1: Characteristics of the three groups.

Group Age-M (yrs) Height-M (cm)UPDRS-M

(score)Gender

PD TG 63.86 (8.41) 170.97 (10.29) 23.68 (10.1)8 male,

6 female

PD OG 73.93 (6.53) 170.72 (10.22) 22.07 (8.0)12 male,2 female

PD CG 67.43 (9.26) 170.15 (6.83) 24.21 (9.5)11 male,3 female

Note: M denotes mean, standard deviations found in brackets.

inputted by the student investigator prior to commencementof training.

A posttest was administered six weeks after the pretest,followed by a six week retention test. During the retentionperiod, participants were told to exercise no more than usual.

2.4. Statistical Analysis. Long-term effects compared mea-surements across time placing pre-, post-, and retention-test values in the same analysis of variance. The dependentvariables analyzed were TUG times, 30-second chair stand,UPDRS III score, and all GAITRite measures. Step lengthand step time data were further analyzed according tomore affected versus less affected lower limb. More affectedlower limb was defined by summing left and right scoresfor question 27 and 28 of the UPDRS III (leg agility andleg tremor, resp.) and taking the greater score. However,after finding no differences, left and right limb data wasautomatically pooled by the statistical analysis software.Also, first and last walking trials of all GAITRite measureswere taken out of the analysis to avoid any learning andfatigue effects. Analysis was carried out by STATISTICA8.0 using a group (treadmill, overground, control) by time(pretest, posttest, and retention test) ANOVA. An alphalevel of 0.05 was used in all analyses. A Tukey’s honestsignificant difference (HSD) post hoc was further employedto determine from where the significant differences weredriven.

3. Results

3.1. Baseline Comparisons. Baseline characteristics can beseen in Table 1. Although the OG group appears to be slightlyolder than TG and CG, one-way ANOVA’s were conductedcomparing all three groups for severity using the UPDRSSection 3 height, initial velocity, step length, and TUG timesresulting in no significant differences (P = 0.81, P = 0.97,P = 0.32, P = 0.20, P = 0.16, resp.).

3.2. Outcome Measures (For summary see Table 2). Steplength showed an overall group by time interaction (F(4,72) =4.5338, P < 0.003), post hoc analysis confirmed thatboth intervention group improved and maintained (afterthe retention period), whereas the control group showedno changes over time (Figure 1). Gait velocity also showedan overall group by time interaction (F(4,72) = 3.7605,P < 0.008), with the interaction being driven by a velocity


Table 2: Mean (6 standard deviation) of outcome measures from pre-, post-, and retention test.

Measure Test PD control PD treadmill group PD overground group ANOVA Pre-, Post-, and Retention

Step length (cm)Pretest 57.7 (12.3) 63.9 (10.6) 57.6 (62.3)

P = 0.003Posttest 59.3 (12.7) 69.4 (9.9)∗∗ 62.3 (8.3)∗∗

Retention test 58.8 (14.0) 69.9 (12.4)∗∗ 64.2 (10.3)∗∗

Velocity (cm/sec)Pretest 109.0 (27.7) 119.2 (15.6) 108.5 (23.8)

P = 0.008Posttest 109.6 (27.1) 128.3 (16.5) 112.2 (18.1)

Retention test 104.3 (32.8) 129.1 (18.0) 118.9 (19.0)

TUG time (seconds)Pretest 9.0 (3.0) 7.7 (2.0) 9.9 (4.2)

P = 0.046Posttest 9.1 (3.3) 6.3 (2.0) 8.4 (3.7)

Retention test 9.1 (3.7) 6.5 (2.5) 10.2 (5.8)

UPDRS scorePretest 24.6 (9.7) 23.6 (10.5) 22.1 (8.0)

NSPosttest 26.7 (8.8) 23.0 (8.0) 25.5 (7.0)

Retention test 26.8 (8.8) 22.6 (8.0) 27.8 (9.1)

NS: denotes a nonsignificant interaction.∗∗: denotes significantly different from pretest (P < .05).3 patients removed from the current analysis.

0

10

20

30

40

50

60

70

80

0 6 12

Step

len

gth

(cm

)

Time (weeks)

TreadmillControlOverground

∗ ∗

Figure 1: Step length significantly improves in TG and OG after sixweeks (posttest) and is maintained after 12 weeks (retention test).

improvement in both training groups but not the controlgroup. Figure 2 displays an overall ∼10 cm/s increase in bothTG and OG, while the CG decreased in gait speed; however,post hoc analysis revealed this change in velocity was notsignificant. There was no change seen in cadence and 30-second chair stand in all groups, across all three testingperiods (P > 0.05). Examination of the TUG test revealeda significant group by time interaction (F(2,39) = 4.0477,P < 0.05), suggesting that only the OG had decreased TUGtimes after the six week intervention. However, while still asignificant interaction (F(4,72) = 2.5564, P < 0.05), after threeparticipants (two in TG, one in CG) were excluded from theanalysis due to medical conditions at the retention period,improvements in TUG time for the OG returned to baselinevalues after time of retention (Figure 3). The UPDRSseverity scores were analyzed and approached a significantinteraction (P = 0.06) (Figure 4). The TG showed a trendto decrease symptom severity from pre- to posttest, and

0

20

40

60

80

100

120

140

160

0 6 12

Vel

ocit

y(c

m/s

)

Time (weeks)


Figure 2: Velocity increases ∼10 cm/s after 12-weeks (retentiontest) in only TG and OG.

improvements were maintained over the retention period.Contrarily, the CG and OG showed a modest symptomseverity increase (i.e., symptoms worsened) from pre- toposttest, which was also maintained after the retention test.All other spatial and timing gait parameters showed nochange.

4. Discussion

While many studies have demonstrated the positive benefitsassociated with visually cued walking in PD, little to nostudies have evaluated long term benefits of visually cued gaittraining. Here we present (according to “level of evidence”and “grading of evidence guidelines”) a Silver BIIa evidencestudy to evaluate the influence of long-term visual cuetraining. A primary objective of the current study was toisolate visual cues in a static versus dynamic context in orderto understand the extent to which optic flow contributes


0

2

4

6

8

10

12

0 6 12

TU

Gti

me

(s)

Time (weeks)


∗

Figure 3: Examination of TUG times reveals a short-term maineffect in PD OG.

to gait improvements in PD. The two gait interventionswere conducted in nearly identical fashions, with the onlydifference between group training protocols being whetherthe cues were on the treadmill (TG) or on the ground (OG).In order to remove any other potential confounds, all othervariables such as intensity, required step length, frequencyof training, and duration of training were kept identicalbetween groups.

Parkinsonian gait has previously been theorized to be theresult of a deficient connection between the basal gangliaand supplementary motor area (SMA). The interactionsbetween these two structures are commonly associated withcontrolling well-learned movements. However, in PD, thisdisconnect is believed to cause impaired internal cueingwithin the basal ganglia, often manifesting itself intoproblematic walking. Visual cues are proposed to bypassthis deficient loop and use visual motor pathways in thelateral premotor cortex (PMC) and posterior parietal cortex(PPC), as these areas are activated through externally cuedmovements and paradoxical movements, respectively [21].Similar gait results in both training groups is evidence thatusual optic flow is not essential, but rather, the transverselines may be activating these areas regardless of surroundingenvironmental information.

It is important to acknowledge however, that we did notcompletely remove optic flow in the treadmill training group.Rather we were able to reduce the amount of optic flowavailable in the treadmill group relative to the overgroundgroup. Thus, some researchers might argue that as long assome optic flow is available, gait improvements can still beachieved.

The findings of our study confirm that transverse lineshave a positive impact on gait parameters [15, 17, 19, 22, 23]and contributes to the existing PD literature on the long-term effects of visual cue training protocols. Step length wasshown to improve after six weeks and was maintained afteran additional six-week retention period in both the TG andOG. Findings indicate that this spatial gait improvement isnot the result of short-term training effects but rather a

0

5

10

15

20

25

30

35

0 6 12

Mot

orsc

ore

Time (weeks)


Figure 4: Motor scores showed at trend in improving symptoms inthe TG, while the OG and CG seem to worsen symptoms.

lasting change in the subjects walking. The fact that steplength improvement was maintained even after the non-exercising period suggests that cueing provides potential forgreater long-term retention gains [24]. The present studyalso rules out any potential strength gains that may havecontributed to step length improvements. The 30-secondchair stand was used as a tool to assess lower limb strengthand gait performance [25] and showed no change across allgroups.

A significant interaction revealed that both interventiongroups achieved faster walking speeds upon completing thecurrent study. Moreover, these same groups also experiencedno change in cadence. In many cueing and treadmill studies,increase in step length is often accompanied by an increasein cadence [17, 26, 27]. However, it is unknown as to whatextent each variable (step length or cadence) influences gaitvelocity. Our study reveals that PD individuals are achievingfaster walking speeds due to taking larger steps rather thanincreasing the frequency of stepping. Thus, their velocityimprovements were a result of step length gains rather than acompensatory reaction to cadence.

Although walking measures were similar across the twointervention groups, the TUG test did reveal an importantdifference. The TUG test has been shown to significantlycorrelate with the Berg balance scale, implying that improve-ment in TUG times may suggest an improvement in dynamicbalance as well [28]. Our study revealed that only the OGsignificantly improved TUG times. This may be due to thenature of the overground walking protocol, which requiredthe participant to turn at the end of the visual cue carpet,mimicking the constant movement found in people walkingon a treadmill. By having individuals turn at the end ofthe 16-meter carpet, the OG may have developed strategiesfor turning. The participants could have used the cues onthe carpet to compete their turn, similar to the way theyuse transverse lines during straight line walking. The linesmay have acted as a critical feedback tool for completinga successful turn, which is essential in optimizing motorperformance [29]. This would suggest that this group may


be simply using an attentional strategy in which the cues areused to focus ones attention on completing the required gaitsequence [17]. Alternatively, it is possible that the OG hadmore opportunity to practise turns relative to the TG (sinceno turns are made on a treadmill). These results providesome important implications for rehabilitation professionals,as turning movements have been problematic in PD andclosely linked to falling incidences and freezing episodes[30, 31]. These effects, however, did not persist past theposttest, which could be due to the complex nature of turningin general compared to straight walking.

In assessing symptom severity, the UPDRS motor scoredisplayed a trend towards a significant interaction (P =0.06). More importantly, there is a hint of symptom improve-ment in the TG, while the OG and CG displayed worseningof symptoms that often accompanies the progression of thedisease [32]. It is important to consider that the improve-ments in the TG may be caused by the treadmill beltdriving proprioceptive inputs [33], when the lower limbsare actively and passively taken through the walking cycle.Impaired proprioception has been previously reported in PDindividuals [34], and perhaps, the treadmill belt is externallystimulating the afferent inputs that may help overcome thesecondary effects of the disease. Hence, future research mightalso consider how external drive may be related to treadmilltraining, while overground training might be more internallydriven. Animal model studies looking at treadmill traininghave been shown to acutely increase dopamine release [35]and chronically upregulate D2 receptors in the striatum ofrats [36]. The motor symptom improvement found in theTG may similarly be the result of this overall availability andutilization of dopamine.

5. Conclusions

The overall improvements found in the treadmill andoverground groups as compared to the control group areindicative of the positive impact transverse lines have ongait. However, similar step length and velocity improvementsin both training groups suggests that typical optic flow isnot necessarily required to achieve short- and/or long-termbenefits associated with PD gait training. Rather than usingoptic flow information, PD participants may be using visionas a strategy to overcome lower limb proprioceptive deficitand/or focus attention on consciously achieving the steppingprocess. Interestingly, hints of motor severity improvementin the TG seem to be driven by additional proprioceptiveinput fed by the belt, while functional tests such as theTUG improved for those that repetitively practised turning.The results of our study reveal that a reduced amountof optic flow can produce similar benefits during gaittraining, and clinically, the implementation of transverselines as a long-term cueing therapy for Parkinson’s diseaseseems appropriate. Furthermore, future work should focuson implementing visual cueing therapy during functionalaspects of walking such as gait initiation, termination, andturning.

Author’s Contributions

Each of the coauthors contributed equally to the develop-ment and completion of this project.

Conflict of Interests

The authors declare no potential conflict of interests withrespect to the authorship and/or publication of this paper.

Acknowledgments

The authors would like to thank Frederico Pieruccini-Fariafor his assistance in this project. This work is supported byan NSERC grant, the Canadian Foundation for Innovationand Sun Life Financial.

References

[1] M. E. Morris, R. Iansek, T. A. Matyas, and J. J. Summers,“Ability to modulate walking cadence remains intact inParkinson’s disease,” Journal of Neurology Neurosurgery andPsychiatry, vol. 57, no. 12, pp. 1532–1534, 1994.

[2] M. Glickstein and J. Stein, “Paradoxical movement in Parkin-son’s disease,” Trends in Neurosciences, vol. 14, no. 11, pp. 480–482, 1991.

[3] C. Assaiante, A. R. Marchand, and B. Amblard, “Discretevisual samples may control locomotor equilibrium and footpositioning in man,” Journal of Motor Behavior, vol. 21, no. 1,pp. 72–91, 1989.

[4] J. J. Gibson, “The visual perception of objective motion andsubjective movement,” Psychological Review, vol. 61, no. 5, pp.304–314, 1954.

[5] J. P. Azulay, S. Mesure, B. Amblard, O. Blin, I. Sangla, and J.Pouget, “Visual control of locomotion in Parkinson’s disease,”Brain, vol. 122, part 1, pp. 111–120, 1999.

[6] T. Prokop, M. Schubert, and W. Berger, “Visual influence onhuman locomotion. Modulation to changes in optic flow,”Experimental Brain Research, vol. 114, no. 1, pp. 63–70, 1997.

[7] A. Lamontagne, J. Fung, B. J. McFadyen, and J. Faubert, “Mod-ulation of walking speed by changing optic flow in personswith stroke,” Journal of NeuroEngineering and Rehabilitation,vol. 4, article 22, 2007.

[8] J. L. Song and J. Hidler, “Biomechanics of overground vs.treadmill walking in healthy individuals,” Journal of AppliedPhysiology, vol. 104, no. 3, pp. 747–755, 2008.

[9] S. Frenkel-Toledo, N. Giladi, C. Peretz, T. Herman, L.Gruendlinger, and J. M. Hausdorff, “Effect of gait speedon gait rhythmicity in Parkinson’s disease: variability ofstride time and swing time respond differently,” Journal ofNeuroEngineering and Rehabilitation, vol. 2, article 23, 2005.

[10] O. Bello, J. A. Sanchez, and M. Fernandez-del-Olmo, “Tread-mill walking in Parkinson’s disease patients: adaptation andgeneralization effect,” Movement Disorders, vol. 23, no. 9, pp.1243–1249, 2008.

[11] M. Suteerawattananon, G. S. Morris, B. R. Etnyre, J. Jankovic,and E. J. Protas, “Effects of visual and auditory cues ongait in individuals with Parkinson’s disease,” Journal of theNeurological Sciences, vol. 219, no. 1-2, pp. 63–69, 2004.

[12] C. A. Lebold and Q. J. Almeida, “An evaluation of mechanismsunderlying the influence of step cues on gait in Parkinson’s


disease,” Journal of Clinical Neuroscience, vol. 18, no. 6, pp.798–802, 2011.

[13] C. A. Lebold and Q. J. Almeida, “Evaluating the contributionsof dynamic flow to freezing of gait in parkinson’s disease,”Parkinson’s Disease, Article ID 732508, 2010.

[14] A. Nieuwboer, “Cueing for freezing of gait in patients withParkinson’s disease: a rehabilitation perspective,” MovementDisorders, vol. 23, supplement 2, pp. S475–S481, 2008.

[15] J. Purdon Martin, “The basal ganglia and postural mecha-nisms,” Agressologie, vol. 18, pp. 75–81, 1977.

[16] Y. Jiang and K. E. Norman, “Effects of visual and auditory cueson gait initiation in people with Parkinson’s disease,” ClinicalRehabilitation, vol. 20, no. 1, pp. 36–45, 2006.

[17] M. E. Morris, R. Iansek, T. A. Matyas, and J. J. Summers,“Stride length regulation in Parkinson’s disease: normalizationstrategies and underlying mechanisms,” Brain, vol. 119, part 2,pp. 551–568, 1996.

[18] E. van Wegen, I. Lim, C. De Goede et al., “The effects of visualrhythms and optic flow on stride patterns of patients withParkinson’s disease,” Parkinsonism and Related Disorders, vol.12, no. 1, pp. 21–27, 2006.

[19] B. Sidaway, J. Anderson, G. Danielson, L. Martin, and G.Smith, “Effects of long-term gait training using visual cues inan individual with Parkinson disease,” Physical Therapy, vol.86, no. 2, pp. 186–194, 2006.

[20] D. Podsiadlo and S. Richardson, “The timed “Up and Go”:a test of basic functional mobility for frail elderly persons,”Journal of the American Geriatrics Society, vol. 39, no. 2, pp.142–148, 1991.

[21] T. Hanakawa, H. Fukuyama, Y. Katsumi, M. Honda, andH. Shibasaki, “Enhanced lateral premotor activity duringparadoxical gait in parkinson’s disease,” Annals of Neurology,vol. 45, no. 3, pp. 329–336, 1999.

[22] R. Marchese et al., “The role of sensory cues in the rehabilita-tion of Parkinsonian patients: a comparison of two physicaltherapy protocols,” Movement Disorders, vol. 15, no. 5, pp.879–883, 2000.

[23] M. Ferrarin, M. Rabuffetti, M. Tettamanti, R. Pignatti, A.Mauro, and G. Albani, “Effect of optical flow versus attentionalstrategy on gait in Parkinson’s disease: a study with a portableoptical stimulating device,” Journal of NeuroEngineering andRehabilitation, vol. 5, article 3, 2008.

[24] T. Platz, R. G. Brown, and C. D. Marsden, “Training improvesthe speed of aimed movements in Parkinson’s disease,” Brain,vol. 121, part 3, pp. 505–514, 1998.

[25] Y. Masuda, Y. Nisida, and K. Kurosawa, “Relationship of a 30-second chair-stand test to gait performance in stroke patients,”Rigakuryoho Kagaku, vol. 19, no. 2, pp. 69–73, 2004.

[26] A. R. Lindquist, C. L. Prado, R. M. L. Barros, R. Mattioli, P.H. Lobo Da Costa, and T. F. Salvini, “Gait training combiningpartial body-weight support, a treadmill, and functionalelectrical stimulation: effects on poststroke gait,” PhysicalTherapy, vol. 87, no. 9, pp. 1144–1154, 2007.

[27] T. Toole, C. G. Maitland, E. Warren, M. F. Hubmann, andL. Panton, “The effects of loading and unloading treadmillwalking on balance, gait, fall risk, and daily function inParkinsonism,” NeuroRehabilitation, vol. 20, no. 4, pp. 307–322, 2005.

[28] S. Bennie et al., “Measurements of balance: comparison of thetimed “Up and Go” test and functional reach test with the bergbalance scale,” Journal of Physical Therapy Science, no. 15, pp.93–97, 2003.

[29] J. Carr and R. B. Shepard, Neurological Rehabilitation: Opti-mizing Motor Performance, The Bath Press, Kent, 1998.


[31] S. Rahman, H. J. Griffin, N. P. Quinn, and M. Jahanshahi, “Thefactors that induce or overcome freezing of gait in Parkinson’sdisease,” Behavioural Neurology, vol. 19, no. 3, pp. 127–136,2008.

[32] G. Alves, T. Wentzel-Larsen, D. Aarsland, and J. P. Larsen,“Progression of motor impairment and disability in Parkinsondisease: a population-based study,” Neurology, vol. 65, no. 9,pp. 1436–1441, 2005.

[33] O. Bello, G. Marquez, M. Camblor, and M. Fernandez-Del-Olmo, “Mechanisms involved in treadmill walking improve-ments in Parkinson’s disease,” Gait and Posture, vol. 32, no. 1,pp. 118–123, 2010.

[34] G. Abbruzzese and A. Berardelli, “Sensorimotor integration inmovement disorders,” Movement Disorders, vol. 18, no. 3, pp.231–240, 2003.

[35] R. Meeusen, M. F. Piacentini, and K. De Meirleir, “Brainmicrodialysis in exercise research,” Sports Medicine, vol. 31, no.14, pp. 965–983, 2001.

[36] P. G. MacRae, W. W. Spirduso, and T. J. Walters, “Endurancetraining effects on striatal D2 dopamine receptor bindingand striatal dopamine metabolites in presenescent older rats,”Psychopharmacology, vol. 92, no. 2, pp. 236–240, 1987.


Research Article

Walking Ability Is a Major Contributor to Fear of Falling inPeople with Parkinson’s Disease: Implications for Rehabilitation

Maria H. Nilsson,1 Gun-Marie Hariz,2, 3 Susanne Iwarsson,1 and Peter Hagell1, 4

1 Department of Health Sciences, Lund University, Box 157, SE-221 00 Lund, Sweden2 Department of Community Medicine and Rehabilitation, Umea University, SE-901 87, Umea, Sweden3 Department of Pharmacology and Clinical Neuroscience, Umea University, SE-901 87, Umea, Sweden4 School of Health and Society, Kristianstad University, SE-291 88 Kristianstad, Sweden

Correspondence should be addressed to Maria H. Nilsson, maria [email protected]

Received 31 May 2011; Accepted 15 July 2011


Copyright © 2012 Maria H. Nilsson et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

Although fear of falling (FOF) is common in people with Parkinson’s disease (PD), there is a lack of research investigating potentialpredictors of FOF. This study explored the impact of motor, nonmotor, and demographic factors as well as complications of drugtherapy on FOF among people with PD. Postal survey data (including the Falls Efficacy Scale, FES) from 154 nondemented peoplewith PD were analyzed using multiple regression analyses. Five significant independent variables were identified explaining 74%of the variance in FES scores. The strongest contributing factor to FOF was walking difficulties (explaining 68%), followed byfatigue, turning hesitations, need for help in daily activities, and motor fluctuations. Exploring specific aspects of walking identifiedthree significant variables explaining 59% of FOF: balance problems, limited ability to climb stairs, and turning hesitations. Theseresults have implications for rehabilitation clinicians and suggest that walking ability is the primary target in order to reduce FOF.Specifically, balance, climbing stairs, and turning seem to be of particular importance.

1. Introduction

People with Parkinson’s disease (PD) have an increased riskof falling [1], and fear of falling (FOF) is also more commonand pronounced compared to controls [2–6]. FOF has beendescribed as an ongoing concern about falling, a loss of bal-ance confidence, low fall-related self-efficacy, or as activityavoidance [7–11].

The prevalence of FOF in people with PD has beenreported to range from 35% to 59% [2, 12–16], although astudy that included only men reported a lower prevalence(18%) [17]. It is even more common and pronounced amongfallers [2, 6, 12–14, 17–19]. FOF can cause social isolation[20], and up to 70% of people with PD report activitylimitations due to FOF [2, 21]. It is thus important for reha-bilitation clinicians to understand the factors contributing toFOF.

Successful interventions need to be based on an under-standing of factors associated with (and potentially influenc-ing) the target of the intervention. That is, if rehabilitationaims to reduce FOF, it should target factors that may influ-ence FOF. With respect to FOF in PD, weak to moderateassociations (Spearman correlations (rs)) have been foundbetween FOF and age (rs, ≤ 0.08), PD duration (rs, <0.29),and disease severity (rs, 0.47) [14, 22]. Previous studieshave also shown that FOF relates to freezing of gait (FOG)[15, 23], physical functioning [14], gait tests [5, 14, 22, 24],balance [3, 14, 22], mobility, activities of daily living (ADL)[14, 21], and sex [14]. However, those studies have reliedon bivariate analyses, and none has simultaneously taken abroader range of independent variables (e.g., motor symp-toms, drug therapy complications such as motor fluctuationsand dyskinesias, nonmotor symptoms, and demographicfactors) into account. The objective of this study was to


explore the potential contributions of motor, nonmotor, anddemographic factors, as well as complications of drug ther-apy, on FOF among people with PD.

2. Participants and Methods

Data were collected by a postal survey to a sample of peo-ple with idiopathic PD [25]. All individuals with PD thatreceived care at a Swedish university hospital were consideredfor inclusion in the study. Exclusion criteria constituteddementia or severe cognitive impairment as determined bytheir respective PD-specialized nurse clinicians. The surveywas sent to 282 individuals (39% women) followed by a re-minder about ten days later. Of 231 returned questionnaires,38 were returned blank and two were returned to senderdue to a change of address. There were thus 191 surveyrespondents (43% women; conservative total response rate,68%). Six of these had left the included FOF-questionnairecompletely blank, and total scores could not be computedfor another 31 participants due to missing data. These 37persons were excluded from the analysis. Excluded partici-pants did not differ (P ≥ 0.153) from those included withrespect to sex, age, and PD duration. Characteristics of thefinal study sample (n = 154) are presented in Table 1. Theinvestigators did not have access to patient details (beyondthose provided by survey responders) or addresses. Thestudy was conducted in accordance with the Declaration ofHelsinki, and all participants gave their informed consent.

2.1. Survey Questions and Instruments. In addition to demo-graphic questions, the survey included a set of questions onthe presence or absence (no/yes) of motor fluctuations (i.e.,a fluctuating effect of anti-PD medications with periods ofmore severe motor symptoms), dyskinesias (i.e., involuntary,irregular, twisting, and/or jerky movements), comorbidity,FOF, falls during the past six months (described and definedas by Lamb et al. [26]), near falls (described and definedas by Gray and Hildebrand [27]), and need of help fromothers in daily activities. Overall perceived PD severity wasself-rated as “mild,” “moderate,” or “severe.” In addition,participants were asked whether they had responded to thesurvey themselves (with or without assistance in readingand/or writing).

A battery of self-administered questionnaires was in-cluded. The Falls Efficacy Scale (FES) conceptualizes FOF aslow fall-related self-efficacy [8]. The Swedish version, FES(S),includes 13 items (activities) rated from 0 (not con-fidentat all) to 10 (completely confident) [14, 28]. The maximumtotal score is 130 points, and a higher score denotes “better”balance confidence. The Functional Assessment of ChronicIllness Therapy-Fatigue scale (FACIT-F) consists of 13 itemswith a total score ranging between 0–52 (higher scores = lessfatigue) [29, 30]. The physical functioning (PF) scale fromthe Short Form-36 (SF-36) includes ten items, and the totalscore can range between 0–100 (higher scores = better) [31,32]. The self-administered version of the FOG Questionnaire(FOGQsa) consists of six items graded 0–4 (higher = worse)[15]. In this study we only used items 3 (freezing: “Do you

feel that your feet get glued to the floor while walking, makinga turn or when trying to initiate walking (freezing)?”) and 6(turning hesitations: “During the past week, how long haveyour typical “freezing” episodes been when making a turn?”) ofthe FOGQsa. Those scoring≥1 on item 3 were categorized as“freezers,” and those scoring≥1 on item 6 were considered tohave turning hesitations. The generic version of the Walk-12(Walk-12G) assesses walking difficulties in everyday life fromthe individual’s perspective [33–35]. The total Walk-12Gscore ranges between 0–42 points (higher scores = worse). Inthis study, item 6 (“Have you had problems balancing whenstanding or walking?”) of the Walk-12G (graded 0–4) wasspecifically used to identify and describe balance problems.Those scoring ≥1 were considered having balance problems.The pain section of the Nottingham Health Profile (NHP-Pain) has eight items and yields a total score between 0–100(higher scores = more pain) [36, 37].

All included patient-reported rating scales have previ-ously been found to have acceptable validity and reliabilityin people with PD [14, 15, 30, 32, 35, 37]. Reliabilities(coefficient alpha) in this study were as follows: FES(S), 0.98;FACIT-F, 0.85; PF, 0.93; Walk-12G, 0.96; NHP-Pain, 0.85.Corrected item-total correlations in this study were all≥0.30.These data support the adequacies of scores used in this study[38].

2.2. Statistical Analysis. Data were checked regarding under-lying assumptions and described and analyzed accordinglyusing PASW version 18 (SPSS Inc., Chicago, IL). The alphalevel of significance was set at 0.05 (2-tailed, exact P-valueswere used).

Spearman correlations (rs) and Mann-Whitney U-testswere used for bivariate analyses of associations with FOF,that is, FES(S). Variables significantly associated with FES(S)scores in bivariate analyses were then entered as independentvariables in regression models with FES(S) scores as thedependent variable. To ease interpretation, all scores wereadjusted to be in the same direction (higher scores = moreproblems) before being entered into the regression analyses.

A first regression model (method: forward) includedmotor, nonmotor, and demographic factors as well as drugtherapy complications (i.e., fluctuations and dyskinesias) asindependent variables. Further details about the includedindependent variables are provided as footnotes in Table 2.Based on results from the first model, a second model wasexplored (method: enter with manual backward deletion)consisting of items from scales found significant in the firstmodel. These items (independent variables) were select-ed based on whether they appeared to represent specificaspects potentially suitable for rehabilitation interventions,in combination with clinical considerations. Details aboutthe included independent variables are provided as footnotesin Table 3.

3. Results

Eighty-five % (131/154) of the participants respondedcompletely independently to the postal survey, whereas the


Table 1: Sample characteristics and bivariate associations with FES(S) scores (n = 154).

Total sample Spearman correlations with FES(S) scores P value

Mean (SD) age, years 70 (9.1) −0.24 0.003

Mean (SD) PD duration, years 6 (5.4) −0.42 <0.001

Fatigue (FACIT-F), median (q1–q3) 36 (27–42) 0.67 <0.001

Physical function (PF), median (q1–q3) 65 (40–84) 0.79 <0.001

Pain (NHP), median (q1–q3) 0 (0–25) −0.50 <0.001

Walk-12G, median (q1–q3) 13 (6–23) −0.82 <0.001

n/total % Median (q1–q3) FES(S) scores P-valuea

Dichotomous variables No Yes

Education: university degree 37/153 24 115 (69–130) 112 (70–130) 0.941

Living alone 38/150 25 119 (80–130) 96 (55–130) 0.125

Comorbidity 77/142 50 107 (60–130) 120 (74–130) 0.271

Motor fluctuations 90/152 58 124 (86–130) 104 (64–128) 0.010

Dyskinesia 57/153 37 124 (85–130) 93 (53–117) <0.001

Freezing of gaitb 57/152 37 128 (112–130) 69 (47–101) <0.001

Turning hesitationsc 58/150 38 128 (113–130) 69 (48–102) <0.001

Experienced falls 50/149 33 123 (90–130) 81 (44–113) <0.001

Experienced near falls 69/147 45 129 (111–130) 84 (52–115) <0.001

Needing help from others in dailyactivities

42/153 27 124 (104–130) 59 (35–91) <0.001

Sex, women 62/152 41Women

116 (52–130)Men

112 (77–130)0.407

Possible score ranges: FACIT-F, 0–52 (higher = better); PF, 0–100 (higher = better); NHP-Pain, 0–100 (higher = worse); Walk-12G, 0–42 (higher = worse);FES(S), 0–130 (higher = better).aMann Whitney U-test.bAs assessed by item 3 (“freezing”) of the FOGQsa (Freezing of Gait Questionnaire, self-administered version). Those scoring ≥1 were categorized as freezers.cAs assessed by item 6 (“turning hesitations”) of the FOGQsa. Those scoring ≥1 were categorized as having turning hesitations.FACIT-F: the Functional Assessment of Chronic Illness Therapy-Fatigue scale; FES(S): the Swedish version of the Falls Efficacy Scale; NHP: the NottinghamHealth Profile; PD: Parkinson’s disease; SD: standard deviation; q1–q3: first and third quartiles.

Table 2: Multiple linear regression with fear of falling (FES(S) scores) as the dependent variable among people with Parkinson’s diseasea,b.

Significant independent variablesc B (95% CI) β P-valueAdjusted R2

Stepwise change Cumulative

Walking difficulties (Walk 12-G) 1.7 (1.2, 2.2) 0.55 <0.001 0.680 0.680

Fatigue (FACIT-F) 0.74 (0.26, 1.2) 0.22 0.003 0.023 0.703

Turning hesitations (item 6, FOGQsa) 11 (2.5, 19.6) 0.15 0.012 0.014 0.717

Need help from others in daily activities 10 (0.96, 19) 0.13 0.030 0.010 0.727

Fluctuations −7.6 (−15, −0.48) −0.11 0.037 0.008 0.735aFor the regression analysis, scores were adjusted to be in the same direction: higher scores = more problems.

bIndependent variables in the analysis were fatigue (FACIT-F), age (years), PD-duration (years), pain (NHP), turning hesitations (item 6, FOGQsa:dichotomized, 1 = turning hesitations), fluctuations (1 = yes), dyskinesia (1 = yes), freezing (item 3, FOGQsa: dichotomized, 1 = freezing), falls (1 = yes), nearfalls (1 = yes), need help from others in daily activities (1 = yes), and walking difficulties (Walk12-G).cListed by order of entry into the model (forward method).B: regression coefficient; CI: confidence interval; β: standardized regression coefficient.

rest attained assistance in reading or writing. The included154 participants had a median FES(S) score of 114 (q1–q3,69–130; min-max, 0–130) and 29% scored at maximum,that is, 130. According to the dichotomous FOF-question,45% (67 out of 149) perceived themselves as having FOF.In addition, 76% (112/148) of the participants experiencedbalance problems when standing or walking. Perceived PDseverity was rated as “moderate” by 96 participants andranged from “mild” (n = 43) to “severe” (n = 14).

Bivariate analyses are presented in Table 1. FES(S) scoresdemonstrated the weakest correlation with age (rs, −0.24)and the strongest (rs, −0.82) with walking difficulties. Thosereporting the presence of motor fluctuations and dyskinesiashad significantly (P ≤ 0.010) lower FES scores (i.e., moreFOF) than those who did not (Table 1). Needing help fromothers in daily activities and experiencing FOG, turninghesitations, prior falls or near falls were also associated withmore (P < 0.001) FOF (Table 1).


Table 3: Explorative multiple linear regression with fear of falling (FES(S) scores) as the dependent variable among people with Parkinson’sdiseasea,b.

Adjusted R2: 0.59B (95% CI) β P-value

Independent variables

Balance problems (item 6, Walk-12G)

Not at all Reference category

A little 3.6 (−6.3, 13) 0.05 0.474

Moderately-extremely 26 (14, 38) 0.36 <0.001

Limited ability to climb stairs (item 5, Walk-12G)

Not at all Reference category

A little 6.8 (−3.4, 17) 0.084 0.188

Moderately-extremely 27 (16, 37) 0.37 <0.001

Turning hesitations (item 6, FOGQsa) 21 (12, 30) 0.29 <0.001aFor the regression analysis, FES(S) scores (range, 0–130) were reversed (0 = better).

bIndependent variables (method: enter with manual backward deletion) were: “Have you been limited in your ability to climb up and down stairs?” (item 5,Walk-12G), “Have you had problems balancing when standing or walking?” (item 6, Walk-12G), “Have you been limited in how far you are able to walk?”(item 11, Walk-12G), turning hesitations (item 6, FOGQsa), and “Has your walking been slow?” (item 12, Walk-12G).The original five response categories of Walk-12G were recoded before being entered in the model: “not at all,” “a little,” or “moderately-quite a bit-extremely.”B, regression coefficient; CI, confidence interval; β, standardized regression coefficient.

In the first regression model, there were signs of mul-ticollinearity between PF and Walk-12G scores (data notshown). PF was therefore omitted from the model infavor of the Walk-12G. This was done because the Walk-12G represents a more specific variable and exhibited asomewhat better reliability than the PF (0.96 versus 0.93).This resulted in a model with five significant independentvariables, explaining 74% of the variance in FES(S) scores(Table 2). The strongest independent variable (as assessedby the standardized regression coefficients, β) was walkingdifficulties, which alone explained 68% of the variancein FES(S) scores. This was followed by fatigue, turninghesitations, needing help from others in daily activities, andmotor fluctuations (Table 2).

In the second explorative regression model, specific gaitand balance items were entered as independent variables(Table 3). In this model, the original five response categoriesof Walk-12G items were recoded and entered as dummyvariables: “not at all” (reference category), “a little,” and“moderately/quiete a bit/extremely,” that is, the three worstcategories were merged into one (due to skewed responsedistributions). Two Walk-12G items were omitted from themodel: item 11 (“Have you been limited in how far you areable to walk?”) due to signs of multicollinearity and item 12(“Has your walking been slow?”) which was not significant.The final model included three significant independentvariables explaining 59% of the variance in FES(S) scores(Table 3). The two strongest independent variables were(moderate to extreme) limitations in climbing stairs andbalance problems. The third significant independent variablewas turning hesitations.

4. Discussion

This study identified that walking disabilities contributedthe strongest to FOF (i.e., low fall-related self-efficacy) in

people with PD. That is, variations in self-rated walkingability could account for a high proportion (68%) of thevariance in FES(S) scores. This is in line with previous studiesshowing a relationship between FOF and clinical gait tests [5,14, 22, 24]. Furthermore, a mixed method pilot study foundthat FOF was universally reported in connection to everydaywalking [39]. Our results have important implications forrehabilitation and suggest that walking difficulties shouldbe the main target in order to reduce FOF. Arguably, suchinterventions may benefit from specifically targeting balanceproblems, stair climbing, and turning hesitations. Theseissues are of particular relevance for the physical therapistwithin the interdisciplinary team.

The present finding of balance problems contributingindependently to FOF is in accordance with previous resultsbased on bivariate analyses [3, 14, 22]. It is noteworthythat prior falls or near falls were not independently asso-ciated to FOF when controlling for the other independentvariables, despite the highly significant bivariate relationshipdemonstrated. This finding illustrates a major pitfall in re-lying on bivariate analyses. However, we did not registerfalls prospectively (as has been recommended [26]), andour sample had a relatively low proportion of fallers. Still,although further confirmatory studies are needed, our find-ings suggest that focus primarily should be put on perceivedbalance impairment rather than on fall prevention per se inorder to reduce FOF.

Impaired balance is common among people with PD,which is confirmed by the fact that 76% of the participantsin our study reported balance problems. This corresponds tothe finding of Schrag et al., who reported that 65% of peo-ple with a PD duration of five years or more experience apostural instability [40]. Although gait and balance trainingare common in rehabilitation for people with PD, veryfew studies have investigated the effects on FOF. Some stu-dies reported improvements after training [41–44], but itis unclear whether these were of clinical significance. In


addition, none of these studies included a long-term fol-lowup and all used different outcome measures, which limittheir comparability. Further studies are therefore warranted,which are of importance since pharmacological treatmentshave insufficient effects on gait and balance problems [45–47]. In addition, although deep brain stimulation in thesubthalamic nuclei has been shown to positively influenceFOF [48, 49], it is a surgical option only eligible for aminority of people with PD.

We identified turning hesitations to be independentlyassociated with FOF. While turning hesitations are related toFOG, it is noteworthy that freezing was not associated withFOF when controlling for the other independent variables inthe identified model, despite the highly significant bivariaterelationship demonstrated between FOF and FOG. Thisfurther illustrates the pitfall in relying on bivariate analyses.The present results suggest that turning hesitations shouldbe more specifically addressed than FOG per se in order toreduce FOF. Turning is in fact impaired in mild PD [50],and rehabilitation clinicians (such as physical therapists)should therefore consider this already early on. Furthermore,moderate to extreme limitations in climbing stairs were alsoindependently associated with FOF, and a previous studyshowed that stairs can cause considerable anxiety amongpeople with PD [39]. This suggests that stair climbing shouldbe considered more specifically both when assessing andtreating people with PD.

In addition to walking difficulties, our primary regressionmodel identified fatigue, need for help in daily activities,and fluctuations as additional but relatively minor factorsassociated with FOF. Although the contributions were small(≤2.3% for each of the variables), this is, as far as we know,the first study showing that fatigue and motor fluctuationsmay be associated with FOF in PD. These results support thevalue of an interdisciplinary approach in the managementof FOF including, for example, an optimization of anti-PDmedications and efforts targeting independence in activitiesof daily living.

There are some methodological concerns associated withthis study. All data were self-reported, and future studiesshould consider including also clinical tests and assessmentsin order to provide a more complete and detailed picture.For example, “having balance problems when standing orwalking” (item 6, Walk-12G) is a coarse indicator of avery complex issue. This item does not take into accountthe complex interaction between the person, environmentand the activity at hand, and it cannot separate balanceproblems in standing from those connected with walking.Although this study considered a relatively broad variety ofaspects, we acknowledge that there may be additional aspectsinfluencing FOF in PD (e.g., cognitive problems, executivedysfunctions, and environmental factors). In addition, oursample was relatively limited and drawn from a universityclinic. It is unknown to what extent such a sample is rep-resentative for the PD population at large, which mayinfluence the external validity of our findings. Finally, theresponse rate of 68% may potentially have introduced abias, particularly since the study design did not allow fora thorough analysis of responders versus nonresponders.

However, excluded responders did not differ from thoseincluded with respect to sex, age, and PD duration, and theprevalence of FOF found here (close to 50%) is in agreementwith that reported in other studies [2, 12–16]. Nevertheless,in order to gain a deeper understanding and reach firmerconclusions, additional quantitative and qualitative work isneeded within this area.

5. Conclusions

This is to our knowledge the first study using multivariateanalysis to explore factors associated with FOF in peoplewith PD. The present results suggest that walking abilityis the primary target in order to reduce FOF. Specifically,balance, climbing stairs, and turning seem to be of particularimportance. Additional studies are warranted in order tofurther improve our understanding of FOF and how to bestapproach it in rehabilitation.

Acknowledgments

The authors wish to thank, Lars Forsgren, Mona Edstromand Birgitta Wikstrom for assistance with patient selectionand data collection. This study was funded by the SwedishParkinson Academy, the Swedish Research Council, the Rib-bing Foundation in Lund, the Swedish Council for WorkingLife and Social Research, and the Faculty of Medicine at LundUniversity. It was accomplished within the BAGADILICO(the Basal Ganglia Disorders Linnaeus Consortium) researchgroup at Lund University, Sweden, and within the context ofthe Centre for Ageing and Supportive Environments (CASE)and the Strategic Research Area Multipark, Lund University,Sweden.

References

[1] S. Deandrea, E. Lucenteforte, F. Bravi, R. Foschi, C. LaVecchia, and E. Negri, “Risk factors for falls in community-dwelling older people: a systematic review and meta-analysis,”Epidemiology, vol. 21, no. 5, pp. 658–668, 2010.



[4] C. Peretz, T. Herman, J. M. Hausdorff, and N. Giladi, “Assesingfear of failing: can a short version of the activities-specificbalance confidence scale be useful?” Movement Disorders, vol.21, no. 12, pp. 2101–2105, 2006.

[5] M. K. Y. Mak and M. Y. C. Pang, “Balance self-efficacy deter-mines walking capacity in people with Parkinson’s disease,”Movement Disorders, vol. 23, no. 13, pp. 1936–1939, 2008.

[6] M. K. Y. Mak and M. Y. C. Pang, “Balance confidence andfunctional mobility are independently associated with falls inpeople with Parkinson’s disease,” Journal of Neurology, vol.256, no. 5, pp. 742–749, 2009.

[7] L. E. Powell and A. M. Myers, “The activities-specific balanceconfidence (ABC) scale,” Journals of Gerontology A, vol. 50, no.1, pp. M28–M34, 1995.


[8] M. E. Tinetti, D. Richman, and L. Powell, “Falls efficacy as ameasure of fear of falling,” Journals of Gerontology, vol. 45, no.6, pp. P239–P243, 1990.

[9] M. E. Tinetti, C. F. Mendes de Leon, J. T. Doucette, and D. I.Baker, “Fear of falling and fall-related efficacy in relationshipto functioning among community-living elders,” Journals ofGerontology, vol. 49, no. 3, pp. M140–M147, 1994.

[10] M. E. Lachman, J. Howland, S. Tennstedt, A. Jette, S. Assmann,and E. W. Peterson, “Fear of falling and activity restriction: thesurvey of activities and fear of falling in the elderly (SAFE),”Journals of Gerontology B, vol. 53, no. 1, pp. P43–P50, 1998.

[11] L. Yardley and H. Smith, “A prospective study of the relation-ship between feared consequences of falling and avoidance ofactivity in community-living older people,” Gerontologist, vol.42, no. 1, pp. 17–23, 2002.

[12] A. C. Dennison, J. V. Noorigian, K. M. Robinson et al., “Fallingin Parkinson disease: identifying and prioritizing risk factorsin recurrent fallers,” American Journal of Physical Medicine andRehabilitation, vol. 86, no. 8, pp. 621–632, 2007.

[13] M. Matinolli, J. T. Korpelainen, R. Korpelainen, K. A.Sotaniemi, V. M. Matinolli, and V. V. Myllyla, “Mobility andbalance in Parkinson’s disease: a population-based study,”European Journal of Neurology, vol. 16, no. 1, pp. 105–111,2009.

[14] M. H. Nilsson, A. M. Drake, and P. Hagell, “Assessment offall-related self-efficacy and activity avoidance in people withParkinson’s disease,” BMC Geriatrics, vol. 10, article 78, 2010.

[15] M. H. Nilsson, G. M. Hariz, K. Wictorin, M. Miller, L.Forsgren, and P. Hagell, “Development and testing of a selfadministered version of the Freezing of Gait Questionnaire,”BMC Neurology, vol. 10, article 85, 2010.

[16] R. Camicioli and S. R. Majumdar, “Relationship betweenmild cognitive impairment and falls in older people with andwithout Parkinson’s disease: 1-year prospective cohort study,”Gait and Posture, vol. 32, no. 1, pp. 87–91, 2010.

[17] K. Robinson, A. Dennison, D. Roalf et al., “Falling risk factorsin Parkinson’s disease,” NeuroRehabilitation, vol. 20, no. 3, pp.169–182, 2005.

[18] M. R. Landers, A. Backlund, J. Davenport, J. Fortune,S. Schuerman, and P. Altenburger, “Postural instability inidiopathic parkinson’s disease: discriminating fallers fromnonfallers based on standardized clinical measures,” Journal ofNeurologic Physical Therapy, vol. 32, no. 2, pp. 56–61, 2008.

[19] M. K. Y. Mak and M. Y. C. Pang, “Fear of falling isindependently associated with recurrent falls in patients withParkinson’s disease: a 1-year prospective study,” Journal ofNeurology, vol. 256, no. 10, pp. 1689–1695, 2009.

[20] W. C. Koller, S. Glatt, B. Vetere-Overfield, and R. Hassanein,“Falls and Parkinson’s disease,” Clinical Neuropharmacology,vol. 12, no. 2, pp. 98–105, 1989.

[21] H. Brozova, J. Stochl, J. Roth, and E. Ruzicka, “Fear of fallinghas greater influence than other aspects of gait disorderson quality of life in patients with Parkinson’s disease,”Neuroendocrinology Letters, vol. 30, no. 4, pp. 453–456, 2009.

[22] F. Franchignoni, E. Martignoni, G. Ferriero, and C. Pasetti,“Balance and fear of falling in Parkinson’s disease,” Parkinson-ism and Related Disorders, vol. 11, no. 7, pp. 427–433, 2005.

[23] A. A. Thomas, J. M. Rogers, M. M. Amick, and J. H. Friedman,“Falls and the falls efficacy scale in Parkinson’s disease,” Journalof Neurology, vol. 257, no. 7, pp. 1124–1128, 2010.

[24] L. Rochester, A. Nieuwboer, K. Baker et al., “Walking speedduring single and dual tasks in Parkinson’s disease: whichcharacteristics are important?” Movement Disorders, vol. 23,no. 16, pp. 2312–2318, 2008.

[25] W. R. G. Gibb and A. J. Lees, “The relevance of the Lewy bodyto the pathogenesis of idiopathic Parkinson’s disease,” Journalof Neurology Neurosurgery and Psychiatry, vol. 51, no. 6, pp.745–752, 1988.

[26] S. E. Lamb, E. C. Jørstad-Stein, K. Hauer, and C. Becker,“Development of a common outcome data set for fall injuryprevention trials: the Prevention of Falls Network Europeconsensus,” Journal of the American Geriatrics Society, vol. 53,no. 9, pp. 1618–1622, 2005.


[28] K. Hellstrom and B. Lindmark, “Fear of falling in patients withstroke: a reliability study,” Clinical Rehabilitation, vol. 13, no.6, pp. 509–517, 1999.

[29] S. B. Yellen, D. F. Cella, K. Webster, C. Blendowski, andE. Kaplan, “Measuring fatigue and other anemia-relatedsymptoms with the Functional Assessment of Cancer Therapy(FACT) measurement system,” Journal of Pain and SymptomManagement, vol. 13, no. 2, pp. 63–74, 1997.

[30] P. Hagell, A. Hoglund, J. Reimer et al., “Measuring fatiguein Parkinson’s disease: a psychometric study of two briefgeneric fatigue questionnaires,” Journal of Pain and SymptomManagement, vol. 32, no. 5, pp. 420–432, 2006.

[31] J. E. Ware Jr. and C. D. Sherbourne, “The MOS 36-item short-form health survey (SF-36). I. Conceptual framework anditem selection,” Medical Care, vol. 30, no. 6, pp. 473–483, 1992.

[32] P. Hagell, A. L. Tornqvist, and J. Hobart, “Testing the SF-36in Parkinson’s disease: implications for reporting rating scaledata,” Journal of Neurology, vol. 255, no. 2, pp. 246–254, 2008.

[33] J. C. Hobart, A. Riazi, D. L. Lamping, R. Fitzpatrick, and A. J.Thompson, “Measuring the impact of MS on walking ability:the 12-item MS Walking Scale (MSWS-12),” Neurology, vol.60, no. 1, pp. 31–36, 2003.

[34] J. Hobart, “The Multiple Sclerosis Impacts Scale (MSIS-29):initial development; subsequent revision; lessons learned,” inQuality of Life Measurement in Neurodegenerative and RelatedConditions, C. Jenkinson, M. Peters, and M. B. Bromberg, Eds.,pp. 24–40, Cambridge University Press, Cambridge, UK, 2011.

[35] S. Bladh, Swedish Walk-12G: validity and reliability in multiplesclerosis and Parkinson’s disease, M.S. thesis, Lund University,Lund, Sweden, 2010.

[36] S. M. Hunt, S. P. McKenna, J. McEwen, E. M. Backett, J.Williams, and E. Papp, “A quantitative approach to perceivedhealth status: a validation study,” Journal of Epidemiology andCommunity Health, vol. 34, no. 4, pp. 281–286, 1980.

[37] P. Hagell, D. Whalley, S. P. McKenna, and O. Lindvall,“Health status measurement in Parkinson’s disease: validityof the PDQ-39 and Nottingham Health Profile,” MovementDisorders, vol. 18, no. 7, pp. 773–783, 2003.

[38] J. Hobart and S. Cano, “Improving the evaluation of ther-apeutic interventions in multiple sclerosis: the role of newpsychometric methods,” Health Technology Assessment, vol. 13,no. 12, pp. 1–177, 2009.


[40] A. Schrag, Y. Ben-Shlomo, and N. Quinn, “How common arecomplications of Parkinson’s disease?” Journal of Neurology,vol. 249, no. 4, pp. 419–423, 2002.

[41] B. D. Cakit, M. Saracoglu, H. Genc, H. R. Erdem, and L. Inan,“The effects of incremental speed-dependent treadmill train-ing on postural instability and fear of falling in Parkinson’s


disease,” Clinical Rehabilitation, vol. 21, no. 8, pp. 698–705,2007.



[44] N. Smania, E. Corato, M. Tinazzi et al., “Effect of balancetraining on postural instability in patients with idiopathicparkinsong’s disease,” Neurorehabilitation and Neural Repair,vol. 24, no. 9, pp. 826–834, 2010.

[45] H. L. Klawans, “Individual manifestations of Parkinson’sdisease after ten or more years of levodopa,” MovementDisorders, vol. 1, no. 3, pp. 187–192, 1986.

[46] A. M. Bonnet, Y. Loria, and M. H. Saint-Hilaire, “Doeslong-term aggravation of Parkinson’s disease result fromnondopaminergic lesions?” Neurology, vol. 37, no. 9, pp. 1539–1542, 1987.

[47] B. R. Bloem, D. J. Beckley, J. Gert Van Dijk, A. H. Zwinderman,M. P. Remler, and R. A. C. Roos, “Influence of dopaminergicmedication on automatic postural responses and balanceimpairment in Parkinson’s disease,” Movement Disorders, vol.11, no. 5, pp. 509–521, 1996.

[48] M. H. Nilsson, P. A. Fransson, G. B. Jarnlo, M. Magnusson,and S. Rehncrona, “The effects of high frequency subthalamicstimulation on balance performance and fear of falling inpatients with Parkinson’s disease,” Journal of NeuroEngineeringand Rehabilitation, vol. 6, no. 1, article 13, 2009.

[49] M. H. Nilsson, S. Rehncrona, and G. -B. Jarnlo, “Fear of fallingand falls in people with Parkinson’s disease treated with deepbrain stimulation in the subthalamic nuclei,” Acta NeurologicaScandinavica, vol. 123, no. 6, pp. 424–429, 2011.



Research Article

Impaired Economy of Gait and Decreased Six-Minute WalkDistance in Parkinson’s Disease

Leslie I. Katzel,1, 2 Frederick M. Ivey,1, 3, 4 John D. Sorkin,1, 2 Richard F. Macko,1, 3, 4

Barbara Smith,5 and Lisa M. Shulman3

1 Baltimore Veterans Affairs Medical Center and Geriatrics Research Education and Clinical Center, Baltimore, MD 20201, USA2 Division of Gerontology & Geriatric Medicine, Department of Medicine, University of Maryland School of Medicine,Baltimore, MD 20201, USA

3 Department of Neurology, University of Maryland School of Medicine, Baltimore, MD 20201, USA4 Maryland Exercise and Robotics Center of Excellence, VA Rehabilitation Research & Development, Baltimore,MD 20201, USA

5 University of Maryland School of Nursing, Baltimore, MD 20201, USA

Correspondence should be addressed to Leslie I. Katzel, [email protected]

Received 10 May 2011; Accepted 5 July 2011


Copyright © 2012 Leslie I. Katzel et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Changes in the biomechanics of gait may alter the energy requirements of walking in Parkinson’s Disease (PD). This studyinvestigated economy of gait during submaximal treadmill walking in 79 subjects with mild to moderate PD and the relationshipbetween gait economy and 6-minute walk distance (6 MW). Oxygen consumption (VO2) at the self-selected treadmill walkingspeed averaged 64% of peak oxygen consumption (VO2 peak). Submaximal VO2 levels exceeded 70% of VO2 peak in 30% ofthe subjects. Overall the mean submaximal VO2 was 51% higher than VO2 levels expected for the speed and grade consistentwith severe impairment in economy of gait. There was an inverse relationship between economy of gait and 6MW (r = −0.31,P < 0.01) and with the self-selected walking speed (r = −0.35, P < 0.01). Thus, the impairment in economy of gait and decreasedphysiologic reserve result in routine walking being performed at a high percentage of VO2 peak.

1. Introduction

Walking capacity is central to the performance of manyactivities of daily living. Difficulty with walking is oneof the cardinal symptoms of Parkinson’s Disease (PD).Alterations in the biomechanics of gait, such as decreasedstride length, increased stride length variability, and reducedgait speed, are common even in early stages of PD [1–3].Most often, PD patients attempt to compensate for shortsteps by increasing gait cadence, thereby potentially alteringenergy requirements. This higher energy cost of movementis often referred to as a lower economy of gait and is afunction of abnormal gait patterns that accompany aging andneurological disability. Reduced economy of gait has beenassociated with impaired function and fatigue in non-PDpopulations [4–9], but there is currently scant informationon how parkinsonian gait affects energy expenditure oreconomy of gait using direct measures of oxygen consump-tion [10]. Further, little is known about the relationship

between economy of gait and mobility. Hence, the purposeof this study was to investigate economy of gait duringsubmaximal treadmill walking in mild to moderate PD, andthe relationship between economy of gait and the distancecovered during the 6-minute walk (6 MW).

2. Methods

2.1. Subjects. Participants for this study were recruited fromthe University of Maryland Parkinson’s Disease Center andthe Baltimore VA Medical Center neurology clinics as partof an exercise intervention trial in PD [11]. Inclusion criteriawere (1) diagnosis of levodopa-responsive PD characterizedby 2 of 3 cardinal signs (resting tremor, bradykinesia,rigidity), (2) Hoehn and Yahr (HY) [12] stage 1 to 3 (while“on” for motor fluctuations), and (3) presence of mild tomoderate gait impairment, (score of 1 or 2 on UnifiedParkinson’s Disease Rating Scale (UPDRS) [13] questions no.


29 Gait or no. 30 Postural Stability, (4) Age ≥ 40, )5) Folsteinmini-mental state examination [14] score ≥ 23, and (6)unlikely to require PD medication adjustment for 4 months.Exclusion criteria were (1) unstable cardiac, pulmonary, liver,or renal disease, (2) unstable hypertension or diabetes, (3)anemia, orthopedic, or chronic pain-restricting exercise, (4)unstable psychiatric illness, or (5) >20 minutes of aerobicexercise more than 3 times per week (to avoid prior trainingeffect). This study was approved by the Institutional ReviewBoard at the University of Maryland, Baltimore, and writteninformed consent was obtained from each participant.

All physical performance measures, rating scales, andfunctional tests were performed while the subjects were “on”or within 3 hours of medication intake. Subjects used anadditional dose of medication to maintain the “on” statewhen necessary.

2.2. Assessments. The UPDRS was administered by a neu-rologist with expertise in movement disorders (LS). TheTotal UPDRS includes three subscales: Mentation, Behavior,and Mood (Part I), Activities of Daily Living (Part II),and the Motor Examination (Part III). Short distanceambulatory function was assessed with three-timed 10 meterwalks. The self-selected walking speed was defined as theaverage velocity of the three tests. This short-distance test iswidely recognized as a valid index of mobility recovery andsimulates the distance required for many home-based dailyfunctions. The 6 MW is a distance that is more representa-tive of community-based daily activities. Participants wereinstructed to cover as much distance as possible in 6 minutes,turning every 100 feet, as prompted by orange traffic conesset apart across a flat, clear space.

2.3. Exercise Treadmill Testing

Screening Treadmill Test. A screening graded-treadmill testto voluntary exhaustion without measurement of the rate ofoxygen consumption (VO2) was performed using a manualprotocol as previously described [15, 16]. All treadmilltesting was performed in the early afternoon while thesubjects were “on”. This screening exercise treadmill testserved to (1) acclimate the subjects to walking on a treadmill(2) evaluate for symptoms of overt coronary disease or todetect silent myocardial ischemia (3) evaluate hemodynamicheart rate and blood pressure response to exercise (4) observegait patterns and (5) determine whether there were anyissues that would preclude their ability to safely exercise. Allsubjects wore a gait belt for safety, and a spotter stood behindsubjects during the treadmill evaluations. Subjects wereinstructed to use the minimum level of handrail support forbalance during the test.

The initial target speed for treadmill testing was thesubject’s self-selected over ground walking velocity, with theincline set at 0%. The first stage was conducted for 2 minutesat 0% grade, the next stage was conducted for 2 minutes at4% grade, and then the grade was subsequently advancedby 2% every minute until voluntary exhaustion. In frailersubjects, the second stage was conducted at 2% instead of

4% for a more gradual increase in workload. Once the gradereached 10%, subjects were asked if the speed of the treadmillcould be simultaneously advanced with grade (generallyby 0.2 mph). The electrocardiogram (ECG) was monitoredcontinuously, and blood pressure was measured during thefirst 3 stages of the tests and every 2 minutes during recovery.

Exercise Treadmill Test with Measurement of Peak OxygenConsumption. At the next study visit one week later, subjectsunderwent a progressive-graded exercise treadmill test tovoluntary exhaustion as described above with measurementof peak oxygen consumption (VO2 peak using a Quark Car-dio Pulmonary Exercise Testing metabolic analyzer (Cosmed,Rome, Italy)). In some subjects, the initial treadmill speedwas adjusted slightly based on the results of the screeningtreadmill test and feedback from the research subjects. Asa result, the average self-selected walking speed on thetreadmill was 94% of their self-selected over ground speed(2.31 ± 0.59 miles per hour (mph) versus 2.46 ± 0.53 mph).The first stage was conducted for 2 minutes at 0% grade(first submaximal treadmill stage), and then advanced asdescribed above. O2 consumption, CO2 production, andminute ventilation were measured breath-by-breath, andvalues averaged for 20 second intervals. Subjects wereinstructed not to talk during the test as this is known toaffect the depth of breathing and gas exchange. Based on ourpilot study [15], we anticipated that we would not be able tomeasure true maximal aerobic capacity (defined as a plateauin oxygen consumption during the final stage, maximal heartrate >85% of age-adjusted predicted maximal heart rate,and respiratory quotient (RQ) or respiratory exchange ratio(RER) > 1.10) in many of these deconditioned subjects. TheVO2 peak was based on the mean of the final two 20-secondaverages obtained during the final stage of the test.

2.4. Economy of Gait. We used the average O2 consumptionvalues obtained over the final 40 seconds of the first sub-maximal treadmill stage to measure economy of gait. The2-minute duration of this stage is similar to the time spenton many activities of daily living. Economy of gait wascalculated as the measured VO2 during the first treadmillstage divided by the predicted VO2 for non-PD age-matchedsubjects based on commonly accepted American College ofSports Medicines equations for subjects walking accountingfor treadmill speed and grade [17].

VO2 = horizontal component + vertical component+ resting component,

VO2 (mL/kg/min) = 0.1 (speed) + 1.8 (speed)(fractional grade) + 3.5,

Speed = speed in meter/minute, to convert to mph,1 mph = 26.8 meter/minute.

Higher oxygen consumption levels for any given speedand treadmill grade imply increased energy expenditure andimpaired economy of gait.

2.5. Statistics. SAS version 9.2 (SAS Institute, Inc, Cary,NC, USA) was used for the statistical analyses. Descriptive


statistics are expressed as mean ± standard deviation (SD).Pearson’s correlation coefficients were used to calculatestrength of relationship between variables. All statistical testswere two sided and performed at a significance level of 0.05.

3. Results

Seventy-nine subjects (57 men and 22 women) completedthis cross-sectional study. Physical characteristics and PDseverity scores are summarized in Table 1. Based on theUDPRS and HY ratings, the subjects had a broad rangeof disease severity from mild to moderately severe PD.Eleven subjects (7%) had received deep brain stimulationsurgery for PD. The level of medical comorbidity in thesample was low, with only five individuals (6%) with priorhistory of stable coronary artery disease, seven (10%) onmedication for diabetes, and only one was a current smoker(1%). Twenty-nine subjects (37%) were on medications forhypertension, including five on betablockers.

The VO2 at the self-selected treadmill walking speedaveraged 64% of their VO2 peak. There were, however, a widerange of values (31% to 89% of VO2 peak). Interestingly,24 of 79 subjects had submaximal VO2 levels that exceeded70% of their VO2 peak, indicating severe reduction ineconomy of gait, with 3 subjects approaching 90% of theirVO2 peak. Overall the subjects had mean submaximal, self-selected walking speed VO2 values that were 51% higherthan the VO2 levels expected for the same speed and gradefor non-PD subjects (13.0 ± 3.3 mL/kg/min versus 9.7 ±1.6 m/kg/min). This observation provides clear evidence ofthe large decreases in economy caused by parkinsonian gaitpatterns (Figure 1).

We examined whether PD severity was associated witheconomy of gait (the ratio of measured VO2 and predictedVO2). There was a significant correlation of HY stagewith economy of gait (Figure 2) with more advanced PDseverity associated with lower economy of gait. There wasno relationship between economy of gait with total or motorUPDRS. There was an inverse relationship between economyof gait and the distance covered during the 6 MW (r =−0.31, P < 0.01). Specifically, individuals whose measuredVO2 was a higher percentage of their VO2 peak during theirself-selected walking speed covered less distance walking forsix minutes (Figure 3). There was also an inverse relationshipbetween walking speed on the treadmill test and economy ofgait (r = −0.35, P < 0.01).

4. Discussion

Our results demonstrate that economy of gait is markedlyimpaired in people with mild to moderate PD that increasesthe energy demands of physical activity. Our subjects walkingat their self-selected pace on the treadmill required onaverage 64% of their VO2 peak. Indeed, 30% of our subjectsused over 70% of their VO2 peak during their self-selectedtreadmill speed, and several subjects approached 90% oftheir VO2 peak. By contrast in healthy younger and olderindividuals, most activities require a small percentage of

Table 1: Subject characteristics, disease severity, and physicalperformance measures.

Parameter (N = 79∗) Mean ± SD Range

Age (years) 65.1 ± 10.7 42 to 86

UPDRS total 47.2 ± 14 15 to 96

UPDRS motor 32.4 ± 10.2 11 to 66

Hoehn and Yahr stage 2.2 ± 0.4 1.5 to 3.0

Hoehn and Yahr Stage 1.5 N = 1 (1%) —




Body mass index (kg/m2) 28.1 ± 4.9 18.0 to 41.6

VO2 peak (mL/kg/min) 22.4 ± 4.8 12.6 to 37.4

Submaximal VO2 (mL/kg/min) 13.0 + 3.3 5.1 to 21.6

Walking speed (mph) 2.31 ± 0.59 1.0 to 3.8

6 min walk distance (meters) 424 ± 106 122 to 695∗

6-min walk performed in 75 subjects.

0

5

10

15

20

25

0 1 2 3 4 5

Speed (mph)

VO

2(m

L/k

g/m

in)

Measured submaximal predicted

Predicted

VO2 versus

Figure 1: Submaximal VO2 measured at self-selected walking speedduring the last 40 seconds of the first 2-minute stage of treadmilltest versus walking speed in mph. Diamonds show measured values,where solid line shows expected value (VO2 mL/kg/min predicted =0.1 × 26.8 speed in mph + 3.5). The vast majority of subjects hadmeasured values higher than the predicted values indicative of pooreconomy of gait.

the maximal or peak working capacity as indexed by theirVO2 peak [18, 19]. In a study of seniors without PD, thepercentage of oxygen uptake (VO2/VO2 peak) during low,moderate and high workload levels was 32%, 42%, and 50%,respectively [20]. In that study, VO2 at the fastest comfortablewalking speed was 40% of VO2 peak, values substantiallylower than those observed in our subjects with mild tomoderate PD.

There was a relationship between HY stage and economyof gait, such that individuals with more severe PD had poorereconomy of gait. Impairments in gait and mobility impact onthe ability of subjects with PD to perform a number of gait-dependent daily activities including housework, dressing,and transferring in and out of bed [21]. Impaired gait


0

0.5

1

1.5

2

2.5

1 1.5 2 2.5 3 3.5

Eco

nom

yof

gait

HY stage

Economy of gait HY stageversus

Figure 2: Relationship between the Hoehn and Yahr stage andeconomy of gait (ratio of measured VO2 to predicted VO2). Highervalues of the ratio of measured VO2 to predicted VO2 are indicativeof impaired economy of gait.

r = −0.31, P < 0.01

0

200

400

600

800

20 40 60 80 100

VO2 at self-selected walking speed of peak (%)

6MW distance and O2 consumption

6M

W(m

)

VO2

Figure 3: There was an inverse relationship between the distancecovered during the 6 min walk (6 MW) and the amount of oxygensubjects consumed at their self-selected walking speed during thefirst stage of the treadmill expressed as a percentage of their VO2

peak.

economy may result from many factors including abnormalgait biomechanics and altered spatiotemporal aspects of gaitassociated with PD, that is, slow, short-stepped shufflinggait with decreased stride length, asymmetric arm [22]swing, tremor and rigidity, postural instability, loss ofrange of motion of axial structures, impaired sensorimotorintegration, and so forth, The modest association betweeneconomy of gait and disease severity and economy of gaitand distance covered during the 6MW test also indicatesthat other factors such as balance problems, difficulty withturning, and physical deconditioning contribute to impairedmobility in these subjects [1–3].

Few studies have directly measured walking economyin PD. Christiansen et al. examined walking economy at anumber of walking speeds in subjects with PD comparedto healthy subjects without PD [10]. VO2 was found tobe 6 to 10% higher in people with PD at walking speedsabove 1 mph. We report much greater impairments inwalking economy than Christiansen et al. The VO2 for oursubjects at 2 mph is 12.5 mL/kg/min, whereas Christiansenet al. reported a VO2 of 11 mL/kg/min at this speed. Thisdifference may be explained by greater PD severity in

our population (mean total UPDRS score, 47 versus 32).We used published equations for VO2 rather than directmeasurement in a control population. The predicted VO2

for a given walking speed derived from younger individualsmay underestimate the energy cost of walking in healthyolder adults [19]. Protas et al. [23] also studied submaximaloxygen consumption during steady-state exercise in PD.Exercise performance was assessed using cycle ergometry inPD and non-PD. The PD group was unable to perform thesame level of exercise as rated by maximum power whencompared with the control group, even though the peakVO2 and heart rate were similar. The authors concluded thatthere was poorer exercise efficiency in the PD group thanin controls. Over a range of submaximal cycling intensities,rates of energy expenditure were about 20% higher in PDthan in controls. Thus, our results support previous findingsof reduced economy of gait and exercise efficiency in PD.

There is growing interest in the effects of aging andmedical comorbidities on bioenergetics and their impacton mobility and other measures of physical performance[18, 24]. We have previously reported that subjects with mildto moderate PD have VO2 peak values 20 to 25% lower thanhealthy age-matched controls [16]. This impairment in VO2

peak, in combination with the higher energy demands ofwalking (lower economy of gait), reduces the physiologicreserve in PD. The decreased physiologic reserve and lowerVO2 peak make it more difficult to perform everyday tasks.The higher energy cost of walking necessitates the use ofanaerobic pathways to meet ordinary energy demands, whichmay be associated with fatigue [18, 20]. Clearly, the decreasedphysiological reserve shown in this study has functionalconsequences as evidenced by impaired 6 MW distance andslow self-selected walking speed, particularly in those withmore severe PD. The predicted 6 MW distance for healthysubjects without PD using the equation of Enrichi andSherrill [25] that takes into consideration age, gender, height,and weight was 509 meters compared to the measured 424meters, a difference of 85 meters, or 17% lower in PD. The6 MW distance in our subjects is comparable to the valuesreported by Falvo and Earhart [26] who reported a 6 MWdistance of 394.1 ± 98.4 m in PD patients of similar age toour subjects.

This study has limitations that may result in an under-estimation of the severity of the impairment of economyof gait. (1) The submaximal O2 utilization was measuredby using O2 utilization during the last 40 seconds of thefirst stage of the treadmill test, when subjects walked attheir self-selected speed and 0% grade. We chose this timeas representative of the time period in which our subjectstypically walked. A number of investigators have advocatedmeasuring submaximal O2 for longer periods of time [7,18, 27]. For example, Alexander et al. measured O2 kineticsin frail and non-frail older adults during a 6-minute sub-maximal exercise bout on the treadmill [7]. The BaltimoreLongitudinal Study of Aging employs a 5-minute stage, butthe data from the first 1.5 minutes is discarded [18]. Thisallows for a longer period of time for the subjects to cometo equilibrium and plateau during the bout of submaximalexercise. We recognize that it is possible that some of subjects


did not plateau during the second minute of the exercisedue to a lag in O2 uptake at the start of exercise reflectingimpaired O2 kinetics. However, any error introduced wouldhave biased our measuring less O2 utilization as subjects withdelayed O2 kinetics would take longer to come to equilibrium[7, 27]. Hence, we potentially understated the degree ofinefficiency of our patient sample with respect to economyof gait. (2) Another limitation is that the O2 consumptionduring exercise includes a resting component for the restingmetabolic rate. Indeed this resting component is included inthe American College of Sports Medicine equation [17]. Thisresting component is often measured with the subjects in thesupine position [28], but others have advocated measuring itby having the subject stand for 5 minutes prior to the walkingtest [9] as this allows an examination of the incrementalO2 utilization attributable to the exercise itself. The restingmetabolic rate in subjects with PD might be affected by age-related changes in body composition, sarcopenia, as well asother changes attributable to PD (i.e., resting tremor andmedication effects). Changes in resting metabolic rate in PDmay be clinically significant as a higher resting metabolicrate is associated with increased mortality in older adults[28]. Even if the increased metabolic needs during exerciseare partially explained by an increased resting metabolicrate, the net effect on ambulatory function is the same;more energy is needed for a given level of ambulation. (3)Another potential confound is the use of handrail supportduring this study. Subjects were instructed to walk on thetreadmill with minimal hand support. Subjects varied in theextent to which they used the side rails for balance support.The use of hand support reduces O2 consumption, againleading to a possible underestimate of their O2 utilization(VO2) and subsequent underestimate of the degree ofimpairment of their economy of gait. (4) These measureswere performed with subjects walking on treadmills. Frenkel-Toledo et al. have proposed that treadmill walking may actas an external pacemaker to improve gait variability [29].If gait biomechanics improve on the treadmill, this wouldreduce oxygen utilization and lead to an overestimate of theireconomy of gait. The gait biomechanics and energetics mightbe different in overground walking. (5) Lastly the 6-minutewalk test required subjects to make tight turns around a cone.This might have adversely impacted the distance covered,particularly in subjects that had limited ability to turn, thatis, turning “en bloc”. Future studies employing portablemetabolic systems could be employed to examine economyof gait during overground walking.

There is substantial interest in whether the abnormalitiesin gait and functional performance in PD can be improved bytreadmill exercise training [30–32]. In a pilot study by Pelosinet al. [32], 10 patients with idiopathic PD underwent 4 weeksof treadmill training (30 min, three times a week for 4 weeks).Walking performance (Timed Up and Go, 6-min and 10-mwalking tests) and metabolic function (oxygen uptake andheart and respiratory rate) were evaluated before training,at the end of treatment and after 30 days with two differentgraded exercises (treadmill and cycle ergometer). Trainingsignificantly improved walking performance. Oxygen uptake,and heart and respiratory rates were significantly decreased

only during graded exercise on the treadmill but not on thecycle ergometer consistent with improved economy of gait,but the data are difficult to interpret due to the way they aredisplayed in the paper.

In summary, this study reinforces prior evidence showingimpaired economy of gait in PD that is associated withimpairment of ambulation at both short and long distance.Reduced economy of gait combined with the reducedVO2 peak results in lower physiologic reserve where evencomfortable gait is performed at a high percentage of VO2

peak. Future research should examine the biomechanical andneuromuscular factors that contribute to impaired walkingeconomy in PD. A better understanding of these factors maylead to new approaches to improve functional performanceand quality of life in PD.

Acknowledgments

This work was supported by the Michael J. Fox Founda-tion for Parkinson’s Research, The National Institute onAging (NIA) Claude D. Pepper Older Americans Indepen-dence Center NIH Grant P30-AG02874, VA RehabilitationResearch & Development Maryland Exercise and RoboticsCenter of Excellence, and the Baltimore VA Medical CenterGRECC. The authers also wish to acknowledge the hard workand efforts of Terra Hill, Jessica Hammers, Kate Fisk, BradHennessie and other members of the study team.

References

[1] K. J. Brusse, S. Zimdars, K. R. Zalewski, and T. M. Steffen,“Testing functional performance in people with Parkinsondisease,” Physical Therapy, vol. 85, no. 2, pp. 134–141, 2005.

[2] J. M. Hausdorff, “Gait dynamics in Parkinson’s disease:common and distinct behavior among stride length, gaitvariability, and fractal-like scaling,” Chaos, vol. 19, no. 2,Article ID 026113, pp. 1–14, 2009.


[4] H. Perrault, “Efficiency of movement in health and chronicdisease,” Clinical and Investigative Medicine, vol. 29, no. 2, pp.117–121, 2006.

[5] N. B. Alexander, G. E. Taffet, F. M. Horne et al., “Bedside-to-bench conference: research agenda for idiopathic fatigue andaging,” Journal of the American Geriatrics Society, vol. 58, no.5, pp. 967–975, 2010.

[6] S. J. Lewis, A. J. Barugh, C. A. Greig et al., “Is fatigueafter stroke associated with physical deconditioning? A cross-sectional study in ambulatory stroke survivors,” ArchivesPhysical Medicine and Rehabilitation, vol. 92, no. 2, pp. 295–298, 2011.

[7] N. B. Alexander, D. R. Dengel, R. J. Olson, and K. M.Krajewski, “Oxygen-uptake (VO2) kinetics and functionalmobility performance in impaired older adults,” Journals ofGerontology—Series A Biological Sciences and Medical Sciences,vol. 58, no. 8, pp. 734–739, 2003.

[8] D. Malatesta, D. Simar, Y. Dauvilliers et al., “Energy cost ofwalking and gait instability in healthy 65- and 80-yr-olds,”


Journal of Applied Physiology, vol. 95, no. 6, pp. 2248–2256,2003.

[9] W. M. Fiser, N. P. Hays, S. C. Rogers et al., “Energeticsof walking in elderly people: factors related to gait speed,”Journals of Gerontology—Series A Biological Sciences andMedical Sciences, vol. 65, no. 12, pp. 1332–1337, 2010.

[10] C. L. Christiansen, M. L. Schenkman, K. McFann, P. Wolfe,and W. M. Kohrt, “Walking economy in people with Parkin-son’s disease,” Movement Disorders, vol. 24, no. 10, pp. 1481–1487, 2009.

[11] L. M. Shulman, L. I. Katzel, F. M. Ivey et al., “Exercise and gait-related disability in Parkinson disease,” in Proceedings of the63rd Annual Meeting of the American Academy of Neurology,Honolulu, Hawaii, USA, April, 2011.


[13] S. Fahn and R. L. Elton, “Unified Parkinson’s disease ratingscale,” in Recent Developments in Parkinson’s Disease, S. Fahn,C. D. Marsden, D. B. Calne, and M. Goldstein, Eds., vol. 2, pp.153–164, Macmillan Health Care Information, New York, NY,USA, 1987.

[14] M. F. Folstein, S. E. Folstein, and P. R. McHugh, “’Mini mentalstate: a practical method for grading the cognitive state ofpatients for the clinician,” Journal of Psychiatric Research, vol.12, no. 3, pp. 189–198, 1975.

[15] F. M. Skidmore, S. L. Patterson, L. M. Shulman, J. D. Sorkin,and R. F. Macko, “Pilot safety and feasibility study of treadmillaerobic exercise in Parkinson disease with gait impairment,”Journal of Rehabilitation Research and Development, vol. 45,no. 1, pp. 117–124, 2008.

[16] L. I. Katzel, J. D. Sorkin, R. F. Macko, B. Smith, F. M.Ivey, and L. M. Shulman, “Repeatability of aerobic capacitymeasurements in Parkinson disease,” Medicine Science inSports & Exercise. In press.

[17] “Metabolic calculations,” in ACSM’s Guidelines for ExerciseTesting and Prescription, M. H. Whaley, P. H. Brubaker, and R.M. Otoo, Eds., pp. 286–289, Lipinncott Williams and Wilkins,Philadelphia, Pa, USA, 7th edition, 2006.

[18] J. A. Schrack, E. M. Simonsick, and L. Ferrucci, “The energeticpathway to mobility loss: an emerging new framework forlongitudinal studies on aging,” Journal of the AmericanGeriatrics Society, vol. 58, no. 2, pp. S329–S336, 2010.

[19] D. Malatesta, D. Simar, Y. Dauvilliers et al., “Energy cost ofwalking and gait instability in healthy 65- and 80-yr-olds,”Journal of Applied Physiology, vol. 95, no. 6, pp. 2248–2256,2003.

[20] S. W. Arnett, J. H. Laity, S. K. Agrawal, and M. E. Cress,“Aerobic reserve and physical functional performance in olderadults,” Age and Ageing, vol. 37, no. 4, pp. 384–389, 2008.

[21] L. M. Shulman, A. L. Gruber-Baldini, K. E. Anderson et al.,“The evolution of disability in Parkinson disease,” MovementDisorders, vol. 23, no. 6, pp. 790–796, 2008.


[23] E. J. Protas, R. K. Stanley, J. Jankovic, and B. MacNeill,“Cardiovascular and metabolic responses to upper- and lower-extremity exercise in men with idiopathic Parkinson’s disease,”Physical Therapy, vol. 76, no. 1, pp. 34–40, 1996.

[24] D. M. Wert, J. Brach, S. Perera, and J. M. VanSwearingen, “Gaitbiomechanics, spatial and temporal characteristics, and the

energy cost of walking in older adults with impaired mobility,”Physical Therapy, vol. 90, no. 7, pp. 977–985, 2010.

[25] P. L. Enrichi and D. L. Sherrill, “Reference equations forthe six-minute walk in healthy adults,” American Journal ofRespiratory and Critical Care Medicine, vol. 158, no. 5, part 1,pp. 1384–1387, 1998.

[26] M. J. Falvo and G. M. Earhart, “Reference equation for 6-minute walk in individuals with Parkinson disease,” Journalof Rehabilitation Research and Development, vol. 46, no. 9, pp.1121–1126, 2009.

[27] S. M. Gunn, A. G. Brooks, R. T. Withers et al., “Determiningenergy expenditure during some household and garden tasks,”Medicine and Science in Sports and Exercise, vol. 34, no. 5, pp.895–902, 2002.

[28] C. Ruggiero, E. J. Metter, V. Melenovsky et al., “High basalmetabolic rate is a risk factor for mortality: the BaltimoreLongitudinal Study of Aging,” Journals of Gerontology—SeriesA Biological Sciences and Medical Sciences, vol. 63, no. 7, pp.698–706, 2008.

[29] S. Frenkel-Toledo, N. Giladi, C. Peretz, T. Herman, L. Gru-endlinger, and J. M. Hausdorff, “Treadmill walking as anexternal pacemaker to improve gait rhythm and stability inParkinson’s disease,” Movement Disorders, vol. 20, no. 9, pp.1109–1114, 2005.

[30] J. Mehrholz, R. Friis, J. Kugler, S. Twork, A. Storch, andM. Pohl, “Treadmill training for patients with Parkinson’sdisease,” Cochrane Database of Systematic Reviews, no. 1,Article ID CD007830, pp. 1–33, 2010.

[31] T. Herman, N. Giladi, L. Gruendlinger, and J. M. Hausdorff,“Six weeks of intensive treadmill training improves gait andquality of life in patients with Parkinson’s disease: a pilotstudy,” Archives of Physical Medicine and Rehabilitation, vol. 88,no. 9, pp. 1154–1158, 2007.

[32] E. Pelosin, E. Faelli, F. Lofrano et al., “Effects of treadmilltraining on walking economy in Parkinson’s disease: a pilotstudy,” Neurological Sciences, vol. 30, no. 6, pp. 499–504, 2009.

Documents

Rehabilitation and Parkinson's Disease